CN1306277C - Manipulator for a test head with active compliance - Google Patents

Abstract

An apparatus supports a load such as an electronic test head. A force sensor detects a force received from a load that results from an imbalance of the load to produce a torque about an axis of rotation of the load. The force source provides a counter force relative to the load in response to the force detected by the force sensor. A method of interfacing an electronic test head mounted in a test head manipulator with an electronic device handler is also provided. The docking method includes measuring a magnitude of an imbalance force along or about at least one of a plurality of axes of motion of the test head manipulator and providing a counter force to the imbalance force.

Description

Manipulator for test heads with active compliance

technical field

The present invention relates to methods and apparatus for balancing a test head relative to one or more axes. Also, the present invention provides methods and apparatus for providing a compliant range of motion for a test head relative to one or more axes.

Background technique

In the testing of integrated circuits, chips, and wafers, it is customary to use a system that includes a test head and an instrument that handles the item under test. The processing instrument may be a packaged device handler, wafer inspector, or other device. For simplicity, we will refer to such instruments as "device processors" or simply "processors". The test head is "docked" with the processor. A circuit connection can then be established between the test head and the integrated circuit so that the test head can perform proper testing.

In general, there are two methods of docking, namely, actuator-driven docking and manipulator-driven docking. A technique known as "actuator-driven" docking was first disclosed in US Patent No. 4,589,815 to Smith (hereinafter '815), and some variations of it were later developed and disclosed in US Patent No. 5,654,631 to Ames. Nos. 5,744,974 to Bogden, 5,982,182 to Chiu et al., 6,104,202 to Slocum et al., and 5,821,764 also to Slocum et al. They are incorporated here in their entirety for reference.

In a general sense, a docking system requires an "alignment structure" on one of the two items to be docked that engages with an "alignment receptacle" on the other of the two items. In the '815 patent, guide pins are included as alignment structures, and guide pin receptacles and gussets are included as alignment receptacles. In the three patents issued to Chiu et al. and Slocum et al., alignment in all six degrees of freedom is provided by dynamic coupling that provides surfaces where "no more than two points of contact are collinear" Provides six touchpoints between. In these patents, a "dynamic surface" on one of the two items to be mated acts as an alignment receptacle; and a "dynamic mating surface" on the other of the two items acts as an alignment structure. In the preferred embodiments described in these patents, the ball or spherical surface is the dynamic mating surface or alignment feature and the groove is the dynamic surface or alignment receptacle. As indicated in these patents, many other surface combinations can be used.

The docking assembly described in the '815 patent is similar to the two-point docking assembly partially shown in the Vertical Plane Handler Position in Figure 13 of this application (the figure on the lower left of Figure 13 is partially cut away) 1340. In Figure 13, only half of the components connected to the test head are shown. These two-point docking assemblies all use 2 guide pins 912 and respective mating holes (not shown in FIG. 13 ) and 2 circular cams 910 . When the cam 910 is rotated by the handle 914 attached to the cam 910, the two halves of the dock are dragged along with the guide pin 912 becoming fully inserted into their mating holes (not shown). A cable 915 links the 2 cams 910 so that they rotate synchronously. The cord layout enables the dock to work by applying force to only one or the other of the two handles 914 . Thus, in this case, the handle is the docking actuator.

Butts made by Reid Ashman Manufacturing Company (RAM) [see website and sales literature] are conceptually similar to those described in the '815 patent. However, in RAM docking pieces, linear cams are used instead of circular cams. In addition, the RAM dock utilizes rigid mechanical links and bell cranks instead of cables to synchronize the motion of the cams. The dock is actuated by one or the other of two handles coupled to respective bell cranks.

Powered actuators can be incorporated into the docking member in a variety of ways. For example, a linear actuator as described above could easily be added to transmit docking drive directly to a cable in an '815 docking piece or a mechanical link or wire cam in another docking piece. Linear actuators can be any of several types including electric motors, electric solenoids, or pneumatic tubes.

The docking pieces described in US Patent Nos. 5,654,631 and 5,744,974 utilize guide pins and small holes to align the two halves. However, the docking piece is driven by a vacuum device that pushes the two halves together when evacuated. The two halves stay locked together as long as the vacuum is maintained.

The docking elements disclosed in US Pat. Nos. 5,821,764, 5,982,182 and 6,104,202 utilize dynamic coupling techniques to provide final alignment between the two halves. Guide pins may also be utilized to provide preliminary alignment. The guide pin can be equipped with a detent mechanism which captures the guide pin in its aperture preventing it from escaping. In the '764 and '202 patents, the jaw mechanism exhibits automatic activation; however, in the '182 patent, motor drive means are used for each of the three guide pins. Also, in the '182 patent, the three motors can operate independently, enabling planarization between mating parts. In all of these publications, linear actuators are used to finally pull the two halves together. The disclosed linear actuator is of the pneumatic type.

The purpose of the above discussion is to provide an overview of some applicable actuator-driven docking techniques. It can be seen that the docking member can be actuated by various means.

An alternative method known as "manipulator-driven" docking is disclosed, for example, in US Patent Nos. 5,600,258 and 5,900,737 to Graham et al. This alternative method provides one or more powered and controlled axes ("controlled axes") of the manipulator to position the test head. For example, and as described in the referenced patent, in the Graham patent the vertical, pitch, and roll axes are all controlled axes. Systems are also known which control only a single axis, eg a vertical axis or a tumbling axis. Position sensors are typically used to provide feedback to the controlled axis regarding the position of the test head relative to the device handler/detector. During the docking process, the controller (or manipulator) operates a controlled axis or several controlled axes, firstly brings the test head into the ready state of the docking structure, and then continues to operate the controlled axes to complete the docking. In [Graham et al. patent], the sensor is used by the controller to properly orient the docking surface of the test head towards the docking surface of the device handler/detector (usually coplanar), and after the docking is complete and has been properly Motion is terminated while testing the electrical connection between the head and the device processor/detector. There is no docking piece actuator and, generally, no separate, independent latch mechanism independent of the manipulator shaft. If there is no separate, independent latch, the manipulator shaft must be locked in a position that holds the test head in the fully docked position.

Furthermore, it is not desirable or practical to have the test head balanced in all axes in a manipulator drive system. An unbalanced test head results in unpredictable and unwanted forces that must be overcome by the drive and control mechanism, by the alignment mechanism, and by the structure of the device handler/detector itself.

The docking system is referred to as a "latch docking" system in which a mechanism separate and independent from the manipulator latches the test head to the handler/detector when fully docked. Systems that hold the test head in the fully docked position by locking only the manipulator shaft are referred to as "non-latch docking" systems. Typically, an actuator-driven docking system is a latching docking system, while a manipulator-driven docking system is a non-latching docking system. However, in another case, a combination of the two is also possible.

Where the manipulator is designed for use with a large test head, it is desirable to have the test head substantially free of movement in as many as six axes, or degrees of freedom of motion, to facilitate ease of controlled movement. This is described, for example, in PCT International Patent Application No. US00/00704 entitled "TEST HEAD MANIPULATOR" and "COUNTER BALANCED VERTICAL DOCKING IN A DRIVEN VERTICAL AXIS TEST HEAD MANIPULATOR". MOTION IN ADRIVEN VERTICAL AXIS TEST HEAD MANIPULATOR)" U.S. Provisional Patent Application No. 60/186,196 discusses manual manipulation and safe powering of axes of motion. This is also true in docking/undocking of device handlers/detectors, where motion is determined by, for example, U.S. Patent No. 4,589,815 to Smith, U.S. Patent Nos. 5,821,764 and 6,104,222 to Slocum et al., or The docking system described in US Patent No. 5,982,182 to Chiu et al. provides. This freedom of movement, or "soft motion," is especially important in the process of interfacing the test head with the handler.

The two axes providing motion parallel to the ground (side to side and inside out) and the axis perpendicular to the floor (top to bottom) are shown in FIG. 14 . These three axes include an x-axis (side to side) 1315 , a y-axis (inside to outside) 1325 , and a z-axis (top to bottom) 1335 . Also shown are three pivot axes including pitch (x) 1310 , roll (y) 1320 , and yaw (and yaw) (z) 1330 axes. In two axes (x and y) parallel to the floor, freedom of movement is generally provided by low friction bearings and tracks etc., or in an articulated arm as described in US Patent No. 4,527,942 also to Smith. In the case of a vertical or upper and lower axis (z-axis), it is typical to use a counterweight to achieve a substantially weightless state, in order to provide the required freedom of movement, unless the axis is intentionally locked by the operator. Other techniques known in up-down shafts are the use of a spring mechanism as described in U.S. Patent No. 4,943,020 to Beaucoup et al., and U.S. Patent No. 5,931,048 to Slocum et al., and U.S. Patent No. 5,931,048 to Smith. Pneumatic devices described in Nos. 5,149,029 and 4,705,447. However, US Patent No. 5,949,002 to Alden notes certain difficulties in taking such an approach and suggests the use of a servo control loop that includes a load cell force sensor. However, techniques involving servo control loops with force and position feedback are complex and expensive, and do not provide an easy way for the operator to steer the system in the event of a failure.

In the case of pitch, roll, and yaw rotations, it is known to set the pivot axes so that they pass almost through the center of gravity of the test head and its attached mounting mechanism and rope. This has been achieved in the roll mode manipulator by adding a briquetting counterweight. On cable pivot manipulators, this is done by providing a variety of shims to vary the length of the inner bracket that backs up on the protruding cable pivot manipulator, as described in US Patent No. 5,450,766 to Holt. accomplish.

As shown in Figure 13, it may be beneficial to locate one or more axes of rotation within the test head at or near the actual center of gravity of the test head and cord assembly. Note that within the test head structure, pitch and yaw motions are combined. In FIG. 13 these are shown as pitch axis 1310 (θX of ±5°) and roll axis 1320 (θZ of ±5°). Roll axis 1330 (θY of ±90-95°) is also shown in FIG. 13 .

In the current system, it can be shown that it is desirable to have a pitch axis of approximately ±5° movement, a roll axis of at least ±90° movement (outside the test head), and approximately ±5° yaw movement within the test head needs. This is because, by using, for example, the "CPPJ" (for pitch) and SplitRing Cable Pivot techniques described in U.S. Patent No. 5,450,766 to Holt and U.S. Patent No. 5,608,333 also to Holt, respectively, at the center of gravity or The extent of structures required to implement these axes nearby is significantly smaller than the extent of structures required to implement these axes when they are external to the test head.

It has been proposed to implement this concept by locating ball bearings near the center of gravity of the test head. External adjustments that move the position of the ball bearings in the in-out direction can be equipped so that the test head can be physically positioned such that the pitch axis ( or tumble) shaft balance. The disadvantage of a ball bearing is that it provides all three rotational degrees of freedom simultaneously.

Although desirable, it is very difficult to locate these internal axes at or very close to the true center of gravity of a real test head. The physical act of varying the amount of pin electronics board inside the test head and the length and weight of the test head cord to meet the needs of a particular test setup end user often results in a significant shift in the center of gravity position and, therefore, significant unbalanced forces. Additionally, as the test head moves through its motion envelope, the force exerted by the cord on the test head changes; this causes variable unbalanced forces when a fixed structure system is used. These unbalanced forces prevent desired free movement relative to one or more axes of motion.

Whether the axis of motion is inside or outside the test head, it is often desirable to reposition the axis of motion away from the center of gravity, as is traditionally done for manipulators. An example of this is positioning the tumble axis very close to the DUT interface for very thick test heads, rather than at the center of gravity of the test head which is usually close to the physical center of the test head. If the test head is one meter thick, then placing the roll pivot axis at the center of gravity, and thus near the physical center, implies that the manipulator will require a vertical range of motion (stroke) of at least one meter (100cm). If it were possible to place the roll pivot axis 13 cm inside the DUT surface, then the vertical stroke of the manipulator would be reduced by 74 cm (100 cm - 26 cm), thereby reducing the overall height of the manipulator, or, for a given bearing structure Larger load capacity allows use of longer booms. However, as noted above, moving the center of gravity of the test head away from the known physical center creates unbalanced forces. Also, these unbalanced forces unbalance the test head so that it cannot move freely along all six axes.

Thus, as test heads become larger and more complex, the unbalanced forces in each of the critical axes of motion correspondingly increase. It would be desirable to have a means of neutralizing these unbalanced forces so that the test head can be efficiently manipulated and positioned while providing for the safety of both the operator and the instrument.

Contents of the invention

The present invention relates to apparatus and methods for balancing a test head relative to a shaft so that the test head is substantially weightless relative to the shaft. The invention is applicable to multiple axes of a test head manipulator. Also, the present invention relates to apparatus and methods for providing a compliant range of motion for a test head relative to at least one axis of the test head.

Description of drawings

Figure 1(a) is a perspective view of a balancing unit coupled to a test head according to an exemplary embodiment of the present invention;

Figure 1(b) is a perspective view of a balancing unit coupled with a test head according to another exemplary embodiment of the present invention;

Figure 1(c) is a perspective view of a balancing unit coupled with a test head according to yet another exemplary embodiment of the present invention;

Figure 1(d) is a perspective view of a balancing unit coupled with a test head according to yet another exemplary embodiment of the present invention;

Figure 2(a) is a perspective view of a balancing unit coupled with a test head mounted in a stand according to an exemplary embodiment of the present invention;

Figure 2(b) is a perspective view of a balancing unit coupled with a test head installed in a stand according to another exemplary embodiment of the present invention;

Figure 3(a) is a perspective view of a balancing unit coupled with a test head supported by internal bearings, according to an exemplary embodiment of the present invention;

Figure 3(b) is a perspective view of a balancing unit coupled with a test head supported by internal bearings according to another exemplary embodiment of the present invention;

Figure 4 (a) is a detailed cross-sectional view of a pneumatic cylinder according to an exemplary embodiment of the present invention;

Figure 4(b) is a detailed cross-sectional view of a hydraulic cylinder according to an exemplary embodiment of the present invention;

Figure 5(a) is a perspective view of a compliant drive mechanism including a position sensor and a mechanical lock, according to an exemplary embodiment of the present invention;

Figure 5(b) is a perspective view of a compliant drive mechanism including a position sensor and a hydraulic lock, according to an exemplary embodiment of the present invention;

Figure 5(c) is a perspective view of a compliant drive mechanism including a position sensor and two compliant centering actuators according to an exemplary embodiment of the present invention;

Figure 6(a) is a perspective view of a compliant drive mechanism and a pneumatic balancing unit according to an exemplary embodiment of the present invention;

Figure 6(b) is a perspective view of a compliant drive mechanism and a pneumatic balancing unit according to another exemplary embodiment of the present invention;

7 is a perspective view of a compliant drive mechanism according to yet another exemplary embodiment of the present invention;

Figure 8(a) is a perspective view of a coupled compliant drive mechanism and a pneumatic balancing unit according to an exemplary embodiment of the present invention;

Fig. 8(b) is a perspective view of a coupled compliant drive mechanism and a pneumatic balancing unit according to another exemplary embodiment of the present invention;

Fig. 9(a) is a first compliant driving mechanism coupled with a pneumatic balance unit acting on a first rotating shaft and a first compliant driving mechanism coupled with a pneumatic balancing unit acting on a second rotating shaft according to an exemplary embodiment of the present invention. Two perspective views of the compliant drive mechanism;

Fig. 9(b) shows the first compliant driving mechanism coupled with the pneumatic balance unit acting on the first rotating shaft and the coupling with the pneumatic balancing unit acting on the second rotating shaft according to another exemplary embodiment of the present invention. A perspective view of a second compliant drive mechanism;

Figure 10(a) is a perspective view of a compliant drive mechanism on a vertical axis, according to an exemplary embodiment of the present invention;

Figure 10(b) is a perspective view of a compliant drive mechanism on a vertical axis according to another exemplary embodiment of the present invention;

Figure 10(c) is a perspective view of a compliant drive mechanism on a vertical axis according to yet another exemplary embodiment of the present invention;

Figure 10(d) is a perspective view of a compliant drive mechanism on a vertical axis according to yet another exemplary embodiment of the present invention;

Figure 10(f) is a perspective view of a compliant drive mechanism on a vertical axis according to yet another exemplary embodiment of the present invention;

11 is a perspective view of a test head manipulator with several compliant drive mechanisms according to another exemplary embodiment of the present invention;

Figure 12(a) is a perspective view of a docking device for docking a test head with a processor according to an exemplary embodiment of the present invention;

Figure 12(b) is a perspective view of a docking device coupled to a processor according to an exemplary embodiment of the present invention;

Figure 12(c) is a cross-sectional view of a docking device for docking a test head with a processor according to an exemplary embodiment of the present invention;

Figure 12(d) is another cross-sectional view of a docking device coupled to a processor according to an exemplary embodiment of the present invention;

Fig. 13 is the assembly diagram of existing test head;

Figure 14 is a legend of six motion axes;

15(a)-15(n) are flowcharts according to various exemplary embodiments of the present invention.

Detailed ways

In the following detailed description of several embodiments of the invention, reference is made to the accompanying drawings listed above. It should be understood that the figures are intentionally not drawn to scale; rather, they are drawn to emphasize important features of the invention.

During device testing, it was apparent that the test head remained fully docked. The handler/detector is equipped with a mechanism for automatically positioning each device to be tested in turn relative to the test interface. This creates low frequency mechanical vibrations in the processor/detector coupled to the test head and manipulator by the docking piece. The energy associated with these vibrations must be safely dissipated. For this reason, many users prefer to use the latch's mating piece and leave the manipulator shaft loose during testing. In this case, the vibrations are conveniently absorbed by the movement of the manipulator shaft and the energy is dissipated in their associated friction. On the other hand, if the manipulator shaft is locked securely during the test, then the less elastic parts of the system will mainly absorb the vibration. The electrical contacts in the interface between the test head and the device under test, including, for example, pogo pins and probes, are delicate mechanical structures that absorb such vibrational energy that may damage or reduce their service life. In non-latching symmetric systems, it is desirable to lock some of the manipulator axes to keep the test head docked while testing. In order for one or more shafts to loosen to absorb vibration, it is necessary to keep them balanced and compliant. Therefore, it is beneficial to keep one shaft in a balanced and compliant state while docking the test head.

An additional consideration is that when docking with a device handler or detector, an unbalanced test head will exert all or a portion of its unbalanced forces on the device it is docking with. Typically, such equipment is only designed to support vertical loads. Unbalanced forces are not necessarily vertical and generally have unpredictable magnitude and direction. Unknown unbalanced forces can potentially cause damage to the instrumentation and compromise the automated mechanisms incorporated therein to process wafers and devices. Therefore, it would be great if unbalanced forces could be neutralized throughout the interval where the test head interfaces with the device handler/detector.

Furthermore, in the case of vertical loading on a device handler/detector, many conventional manipulators are counterbalanced, which keeps the test head in a substantially weightless state. This minimizes the vertical loading of the test head on the device processor/detector, thereby simplifying its architectural requirements. Furthermore, for manipulators with hard-driven vertical axes, part of the load may be transferred to the device processor/detector during docking, which increases the structural requirements on them. Therefore, it would be desirable in a hard-driven vertical axis manipulator with vertical compliance to be able to maintain a balanced state the entire time the test head remains docked with the device handler/detector. In the case of a pneumatically based balancing mechanism, an eventual leakage of gas means a loss of balancing force.

Another consideration is that in test head manipulator systems that include pneumatically compliant mechanisms, when the test head is in close proximity to the docking mechanism, it is often necessary to bring the system into equilibrium immediately prior to docking. In some currently known systems, in order to do this, the test head must move freely throughout its full compliance range. As a rule of thumb, during such a balancing process, the test head may suddenly and unexpectedly move a significant distance, generating a large momentum. Such movement potentially causes collisions with docking equipment, resulting in damage to delicate electrical contacts. In addition, operators may also be at risk.

As test heads become larger and more complex, imbalances in critical axes of motion increase accordingly. It is important to have a means of neutralizing these imbalances so that the test head can be manipulated and positioned efficiently and in the safety of both the operator and the instrument.

One technique that has been proposed is to use one or more cylinders, eg, double-acting pneumatic cylinders, to counteract unbalanced forces. Such a cylinder has two gas inlets and a piston fitted between them. The first gas inlet is called the extended inlet and the second gas inlet is called the retracted inlet. The piston moves according to the difference in air pressure on either side of it. A connecting rod attached to the piston is coaxial with the cylinder and passes through one end to impart linear force and motion to the load. If the pressure on the extend input is greater than the pressure on the retract input, the connecting rod extends from the cylinder. Conversely, if the pressure on the retract input is greater than the pressure on the extend input, then the connecting rod retracts into the cylinder. A position sensor may be included to indicate the position of the piston. It will therefore be appreciated that such pistons and appropriate control systems may be used in connection with a particular axis, for example the pitch axis. As the test head is manipulated and docked, the control system will pump air into the extension or retraction inlets as required to eliminate any unbalanced forces and provide compliant motion. It is conceivable that a double-acting pneumatic cylinder would provide a spring-like action; that is, when the differential pressure has been adjusted to counteract the unbalanced force of the shaft, it could cause the test head to move as if it were indeed in equilibrium with a very small force.

This approach is very difficult to implement for several reasons. First, the only feedback mechanism is the position of the piston. Second, seals are used in pneumatic cylinders to minimize gas leakage from one side of the piston to the other and from the point where the connecting rod crosses the end of the cylinder. These seals in the path of motion provide both static and kinetic friction. Typically, pneumatic cylinders exhibit a characteristic known as a "break away force". The breakaway force of a pneumatic cylinder is the force required to overcome static friction and move the piston within the cylinder. Stiction (static resistance) can be considerable and changes with time, use, and temperature. Static resistance tends to be much greater than kinetic or motion friction. Then, to cause the change, the control system adjusts the air pressure differential as needed until at least a change in position (ie, motion) is detected, and then the motion is analyzed and adjusted to the desired result. Due to stiction (static friction), the air pressure differential to initiate motion can be, and often is, significantly greater than the load imbalance combined with the dynamic friction of the pneumatic cylinder. This can lead to highly nonlinear and unstable control problems. Especially with regard to the dynamic nature of the static drag component, robust solutions are difficult to achieve and maintain.

Double-acting pneumatic cylinders offer the possibility of a useful measure of compensation for imbalances in the test head shaft; however, new approaches are required to achieve useful solutions. A second object of the present invention is to provide measures to overcome the disadvantages of the double-acting pneumatic cylinder in order to utilize its advantages firmly.

In an exemplary embodiment, the subject invention compensates for unbalanced forces in at least one test head shaft so that the test head can be efficiently and safely maneuvered, docked, and undocked. The invention comprises a force source like at least one double-acting pneumatic cylinder (for example) as described above combined with other components and a new basic control scheme. The combination of the cylinder and these components is called a "balancing unit". In summary, for a single shaft, first the test head is set and locked in the desired position relative to its supporting structure. Therefore, the test head is locked relative to the axis under consideration. A force sensor (ie, a bi-directional load cell) included in the balancing unit measures the unbalanced force relative to the locking shaft to determine if there is a significant amount of unbalance and the direction of the unbalance. The air pressure on either side of the double-acting cylinder is then adjusted to reduce the force sensed by the force transducer below some predetermined minimum, depending on the weight of the test head and the characteristics of the cylinder, 5 to 25 pounds is generally considered reasonable. Then, the locker is released, and the pressure difference in the pneumatic cylinder transmits force to the test head through the connecting rod to counteract the unbalanced force. This process can be repeated as often as desired, keeping the selected test head shaft in a balanced, free-to-run state.

Figure 1(a) is a perspective view of a test head balancing system for a test head manipulator according to an exemplary embodiment of the present invention. There is provided a balancing unit 110 having two ends. As mentioned above, the balancing unit is the force source (ie double-acting pneumatic cylinder) used to balance the test head relative to an axis, combined with other components and a new control scheme. A first end of the balancing unit 1101 is connected to the test head 100 . A second end of the balancing unit 110 is connected to a bearing structure (not shown) for the pivot shaft 102 . The purpose of the balance unit 110 is to facilitate the rotation of the test head 100 around the pivot axis 102 when the center of gravity of the test head 100 does not coincide with the pivot axis 102 . In the exemplary embodiment shown in FIG. 1( a ), the balancing unit 110 includes a double-acting pneumatic cylinder, and its associated components, as described below.

The balance unit 110 includes a force sensor 120 and a force bar 112 . Force sensor 110 is coupled to force rod 112 so that it measures force along the force rod. The force bar 112 is connected with the test head 100 through the bearing 116a. If the center of gravity of the test head 100 does not coincide with the pivot axis 102 , then a force is exerted by the test head 100 about the pivot axis 102 . Via its connection to force rod 112 , force sensor 120 measures at least one component of this force.

Force sensor 120 may be a typical bi-directional load cell that may measure and indicate (eg, to a controller) the magnitude and direction of an unbalanced force of test head 100 relative to pivot axis 102 . The force sensor 120 may comprise strain gauges in a well-known manner in a bridge circuit providing a voltage output that varies monotonically with the measured force. Using an analog-to-digital converter and a processor, it can be determined whether the force along the force bar 112 is greater than the maximum tolerable for free motion; and, if so, the direction of the force. Additionally, an analog comparator circuit may be used in a known manner to generate a go/nogo signal indicative of the presence and direction of a significant unbalanced force.

The force rod 112 is slidably attached to a pneumatic cylinder 128 by a lock 118 which includes a lock inlet 126 which actuates the lock 118 and which allows the force rod to move relative to the cylinder when the lock 118 is not actuated. 128 parallel movements. Thus, test head 100 may rotate about axis 102 when lock 118 is not actuated. When the lock 118 is actuated, the force rod 112 can no longer slide relative to the cylinder 128 ; and the test head 100 is locked in a certain position relative to the shaft 102 . Locker 118 is one of several types well known in the art. Depending on the type of locker 118 selected, it may be controlled by an electrical signal, gas input, or other means appropriate to the particular application.

The pneumatic cylinder 128 is simply a source of force for pushing toward or away from the test head 100 . When an unbalanced force is detected by the force sensor 120 , the pneumatic cylinder 128 is used to exert a counter force on the test head 100 through the connecting rod 114 . In this exemplary embodiment, a pneumatic cylinder 128 is utilized which requires a gas system. It is envisioned that counterforce can be achieved by various other means, such as hydraulic cylinders or electromagnetic devices.

The pneumatic cylinder 128 is connected with the connecting rod 114 through a piston 130 . The connecting rod 114 is connected with the test head 100 through the bearing 116b. The connecting rod 114 is arranged parallel to the force rod 112 . The axis of the cylinder 128 and the connecting rod 114 are normal to the axis of rotation 102 of the test head 100 . The axes are arranged such that a force acting along each axis will generate a moment or moment about the test head axis of rotation 102 . Telescoping link 114 with sufficient force will cause test head 100 to rotate about axis 102 .

Pneumatic cylinder 128 is connected by bearings 138 to a support structure (not shown). Inside the pneumatic cylinder 128 , the piston 130 moves according to the difference in air pressure on both sides of the piston 130 . One side of the piston 130 within the pneumatic cylinder 128 includes a protruding inlet 134 . The other side of the piston 130 within the pneumatic cylinder 128 contains a retraction inlet 132 . The gas inlets 132 and 134 of the pneumatic cylinder are respectively connected to relatively high-pressure gas supply sources (not shown) through electrically operated control valves (not shown). Optionally, an accumulator (not shown) may be attached to each inlet 132 and 134 to provide a large amount of gas to act on the piston 130 .

In operation, force sensor 120 is used to detect an unbalanced force from test head 100 ; gas is then provided to retract inlet 132 and extend inlet 134 to counteract the force detected by force sensor 120 . Therefore, a pressure differential is created across the piston 130 . For this pressure differential to be of sufficient magnitude and direction, the objective is to reduce the force in the force bar 112 to a magnitude less than a predetermined maximum allowable unbalanced force.

Still referring to FIG. 1( a ), the pneumatic cylinder 128 also includes two piston position sensors 136a and 136b that indicate the position of the piston 130 within the pneumatic cylinder 128 relative to, for example, a controller. For example, limit switches may be used in a known manner to indicate whether the piston 130 is in the center position, or which end of the pneumatic cylinder 128 it is in. If necessary, more complex position sensing devices can be equipped. For example, potentiometers, absolute encoders, incremental encoders with appropriate electronics, etc. may be used to provide precise positional information relative to the position of the piston 130 and connecting rod 114 relative to the pneumatic cylinder 128 . Preferably, the position sensing mechanism is calibratable and adjustable to facilitate system configuration, setup, and maintenance.

A controller (not shown) may be provided to operate inlet valves 132 and 134 and lock 118 and receive feedback signals from position sensors 136a and 136b and force sensor 120 . For example, upon start-up, the pneumatic cylinder 128 is typically not pressurized for a test head 100 that is not docked with a handler. Before any control action, care should be taken to ensure that the test head 100 and the apparatus are not resting on any foreign objects or structures that may interfere with the upcoming balancing operation. The test head 100 is now in a relatively balanced state, or, more likely, an unbalanced state. Then, the lock 118 is actuated, if not already activated, to lock the test head 100 in a certain position. In some cases, it may be necessary to first move test head 100 to a desired position. This movement may be accomplished by manual means, or, alternatively, the controller may be equipped with appropriate algorithms so that it may utilize the pneumatic cylinder 128 in conjunction with the position feedback. It should be noted that such a control algorithm may be difficult to implement for the aforementioned stiction. This is also true if an attempt is made to stop the test head 100 in the desired position before the lock 118 is applied. However, suitable locks 118 may also be used as brakes by applying locks 118 to stop movement as test head 100 is moved to a desired position.

Now, with the test head 100 locked in a certain position, the controller responds to the signal from the force sensor 120 . The force sensor 120 detects the presence of a significant unbalanced force from the test head 100 about the pivot axis 102 . The force sensor 120 is capable of detecting the magnitude and direction of force. If a significant unbalanced force is detected, then the air pressures in the retraction inlet 132 and the extension inlet 134 are adjusted in order to create a pressure differential across the piston 130 . This pressure differential across piston 130 is of sufficient magnitude to reduce the force measured by force sensor 120 to less than a predetermined maximum allowable unbalanced force. When this is done, the lock 118 is disengaged and the test head 100 is substantially weightless with respect to the pivot axis 102 . The test head 100 can now be rotated about the pivot axis 102 . This process can be repeated as necessary to maintain the test head 100 in equilibrium with respect to the pivot axis 102 .

As the test head 100 rotates, the piston 130 will also move. Like a mechanical spring, the differential pressure increases monotonically with piston displacement. The aerodynamic force exerted on a given side of the piston 130 increases inversely proportional to the volume. However, for displacements where the change in gas volume is relatively small, the equivalent elastic force varies approximately linearly with displacement, and the equivalent "spring constant" is K. That is, F=Kx, where F is the change in force and x is the change in piston position. Preferably, try to keep K small so that the force does not vary considerably over the range of motion. Part K is determined by the change in pressure of the total gas volume with respect to each increment of piston displacement, that is, by V and dP/dx, where V is volume and P is pressure. As the piston 130 approaches one end of the stroke, V becomes small and dP/dx becomes exponentially increasing, so K increases like F. As mentioned above, an accumulator (not shown in Figure 1(a)) may be added. Accumulators are used to significantly increase the available gas volume V while limiting dV/dx and dP/dx to acceptable values. This achieves the so-called "soft spring" effect, which is clearly desired.

Ideally, the above loop should be done only once and not repeated until some change is made to the system. However, the system of pneumatic cylinders 128 is prone to gas leaks, so the above cycle is repeated periodically as needed. Experience has shown that the system will hold sufficient gas for up to about 10 minutes or more before the cycle needs to be repeated. In typical operation, this cycle repeats automatically every few minutes (say, 5 to 8 minutes). Furthermore, in typical operation, it is not necessary to repeat the steps of moving test head 100 to a desired location. Unless the test head 100 has encountered an unexpected hard obstacle or some other accident, it should remain within its permissible range of motion. Indeed, if the test head 100 is now in the position required to perform the desired function, then, It is highly undesirable to have the test head 100 move back to a predetermined desired position.

Therefore, as shown in FIG. 15(a), in order, the operation sequence in the exemplary embodiment is as follows:

1. As shown in step 1501, lock the locker 118.

2. As shown in steps 1502 and 1503, adjust the air inlet pressure until the force in the force bar 112 is less than a predetermined value.

3. As shown in step 1504, the lock 118 is disengaged.

4. As shown in step 1505, go to step 1501 after a time shorter than the time to lose a significant amount of gas from the cylinder 128.

This sequence causes test head 100 to behave as if shaft 102 is passing directly through its center of gravity; we will refer to this as the "normal sequence".

In the exemplary system, the time required for steps 1501, 1502, 1503 and 1504 is less than 4 seconds. This sequence of steps will be referred to as the "force elimination sequence".

Still referring to FIG. 1( a ), two stop rings 122 and 124 or "motion limiters" are attached to the lever 112 with the lock 130 between them. The stop rings 122 and 124 are positioned such that the stop rings 122 and 124 are equidistant from the center of the locker 118 when the piston 130 in the pneumatic cylinder 128 is centered. The distance A between the two stop rings 122 and 124 is some reasonable value less than the total stroke of the piston 130 in the pneumatic cylinder 128 . Thus, stop rings 122 and 124 prevent piston 130 from running to the end of cylinder 128 and prevent piston 130 from unloading force rod 112 and force sensor 120, thereby distorting the process. They simultaneously serve as positive motion stops for the test head 100 . It should be noted that if the system is in an unbalanced state and the locker 118 is disengaged, then one of the two stop rings 122 or 124 must rest on the locker 118 and the force bar 112 and force sensor pick up and measure the unbalanced force .

Other control functions not shown in Fig. 1(a) can be added to the above system. For example, a button may be provided to enable the operator to initiate the control sequence beginning with step 1501 or step 1502 listed above. An indicator light could be added to alert the operator when force bar 112 is locked; this would help the operator avoid attempting to move the manipulator along axis 102 and/or attempting to dock while step 1502 is being performed. Such an attempt would create additional unbalanced forces on the test head 100, and the system would erroneously try to compensate for them. It is also possible to incorporate the system with sensors in the docking mechanism, for example, so that when docked, the system becomes locked and/or disabled. In addition, it is preferable to detect whether the test head 100 is in a ready-to-dock position and to allow the sequence to be performed just prior to docking. This will ensure that the sequence is not carried out when docking actually occurs, and also ensures that unbalanced forces are optimally compensated before docking begins.

Another control loop, which will be referred to as a "lock sequence", is possible and useful in selected applications. During the locking sequence, the test head 100 is maintained in a substantially centered position relative to the shaft 102 when the test head 100 is moved away from the device (not shown in FIG. 1( a )) with which it is to be interfaced. This is the "desired lock position". When the test head 112 is maneuvered into the ready-to-dock position, the force relief sequence is invoked; that is, the air pressure in the pneumatic cylinder 128 is adjusted until the force in the force rod 112 is effectively eliminated. Then, disengage the locking device 118, just can carry out docking. The test head 100 can be placed in the desired locked position either manually or automatically using pneumatic cylinder 128 and position sensors 136a and 136b. That is, at start-up, for an undocked test head 100, the controller will first disable the lock 118. The test head 100 is then moved to a desired locked position relative to the shaft 102 being controlled, for example, approximately centered as indicated by the position sensors 136a and 136b. In the case of automatic control, it is preferable to use the lock 118 as a brake, forcing the test head 100 to stop in the desired locked position, rather than stopping it only by pneumatic action.

When ready to dock using the normal sequence, the test head 100 is maneuvered into the ready-to-dock position with the lock 118 released and the pneumatic cylinder 128 balancing the test head 100 . Depending on the system, the operator or a sensor will provide a ready-to-dock signal when the ready-to-dock position is reached. Based on this signal, a force cancellation sequence is invoked to ensure that the test head 100 is as balanced as possible. When using the lock sequence method, the test head 100 is manipulated into the ready-to-dock position while the lock 118 is engaged; and, when the ready-to-dock signal is received, only steps 1502 and 1503 of the force release sequence listed above are performed in turn. Now, in each case, with the test head 100 ready for docking, just balanced, and free to move in all axes, the docking mechanism is actuated and engaged, pulling the test head 100 into the fully docked position . In some applications, the lock 118 is reapplied when a signal (either by the operator or by a sensor) that it has been fully docked is provided. The latch 118 is still used when testing the integrated circuit with the test head. In both cases, during undocking, the lock 118 remains applied until a "clear-of-dock" signal is provided by an operator or a sensor. In both cases, the force elimination sequence can now be performed. In the case of a locking sequence, the process of placing the test head 100 in the desired center position is performed last, and the lock 118 is engaged. In the case of a normal sequence, the lock 118 is let loose.

As will be described later, in some applications it may be desirable to have the latch 118 disengaged during docking. Further information on docking is provided later.

Figure 1(b) shows the same system as described with reference to Figure 1(a), except that in Figure 1(b), bearings 116a and 116b are coupled together and connected to test head 100 at a single point outside. In contrast, in FIG. 1( a ), bearings 116 a and 116 b are connected to test head 100 , respectively.

The balancing mechanism thus described is applicable to Figs. 6(a), 6(b) and 8(a) etc. The discussion associated with these figures provides more information and details on control gases and other matters.

FIG. 1(c) shows the same system as that shown with reference to FIG. 1(a), except that in FIG. Like the system shown in FIG. 1(a), the system in FIG. 1(c) can implement a normal sequence as described above or a locking sequence as described above, except that a hydraulic locker 140 including pressure sensors 146a and 146b is used. . FIG. 4( b ) provides a more detailed cross-sectional view of the hydraulic locking unit 140 . Here, the force rod 112 is attached to a piston 144 of a double-acting hydraulic cylinder 140 .

As shown in FIG. 1( c ), hydraulic cylinder 140 is rigidly attached to pneumatic cylinder 128 . Hydraulic cylinder 140 is filled with a suitable incompressible fluid (not shown). Inside the hydraulic cylinder 140 , a piston 144 moves according to the pressure difference on both sides of the piston 144 . Each side of the piston 144 includes fluid inlets 154 interconnected via a conduit 152 through the control valve 142 . On one side of the control valve 142 are pressure transducers 146a and 146b , each connected to one of the two fluid inlets 154 .

Since hydraulic fluid is substantially incompressible, when the control valve 142 is closed, the force rod 112 is locked. When the control valve 142 is opened, the force rod 112 is free to move. Movement is impeded only by fluid flow through the system. This flow can be controlled by hose, conduit, orifice size, and valve operation. Proper control of flow can create system damping that can be beneficial. Although the pressure on both sides of the piston 144 is the same at the moment the control valve 142 is closed, the unbalanced force exerted by the test head 100 on the force rod 112 increases the pressure on one side of the piston 144 and increases the pressure on the other side of the piston 144. The pressure on the side is reduced. It follows that the pressure difference determined by the two pressure sensors 146a and 146b is indicative of the magnitude and direction of the force acting on the force rod 112 .

The operation of the system shown in Figure 1(c) is as described for the system of Figure 1(a), except the hydraulic valve 142 is actuated instead of the mechanical lock 118, and the pressure sensor is read and processed by the controller output to indicate force failure in the force bar 112.

During operation, pressure sensors 146a and 146b are used to indicate unbalanced forces from test head 100; air is provided to retract inlet 132 and extend inlet 134 of pneumatic cylinder 128 to counteract the forces sensed by pressure sensors 146a and 146b. FIG. 4( a ) provides a more detailed cross-sectional view of hydraulic cylinder 128 . A pressure differential is created across piston 130 by air pressure entering retraction inlet 132 and extension inlet 134 . For this pressure differential to be of sufficient magnitude and direction, the goal is to reduce the force in the force bar 112 to a magnitude less than a predetermined maximum allowable unbalanced force.

In FIG. 1( c ), the process of balancing the test head 100 with respect to the pivot axis 102 is as follows. With the control valve 142 open, the test head 100 is in the desired position relative to its support structure (not shown), and the control valve 142 is then closed. At this point, the test head 100 is locked relative to the pivot shaft 102 because the control valve 142 is closed. Pressure sensors 146a and 146b detect the presence of a significant unbalanced force from test head 100 about pivot axis 102, and pressure sensors 146a and 146b are capable of detecting the magnitude and direction of the force. If a significant unbalanced force is detected, the air pressures in the retraction inlet 132 and the extension inlet 134 are adjusted in order to increase the pressure differential across the piston 130 . With respect to the pressure differential across piston 130, the objective is for the pressure differential to be large enough to reduce the force measured by pressure sensors 146a and 146b to a magnitude less than a predetermined maximum allowable unbalanced force. At this point, the control valve 142 is opened and the test head 100 is substantially weightless in the direction of the pivot axis 102 . The test head 100 can now be moved in the direction of the pivot axis 102 . This process may be repeated as necessary to maintain test head 100 in equilibrium relative to pivot axis 102 .

Figure 1(d) shows the same system as that shown in Figure 1(c), except that in Figure 1(d), bearings 116a and 116b are coupled together and connected to test head 100 at a single point outside. In contrast, in FIG. 1(c), bearings 116a and 116b are connected to test head 100, respectively.

Figure 2(a) shows a further exemplary embodiment of the present invention. FIG. 2( a ) is a perspective view showing application of the balance unit 110 to the roll axis (pivot axis) 102 of the test head 100 mounted on a conventional stand 200 . The test head 100 is connected to the bracket 200 at the two ends of the rolling axis 102 . There is provided a balancing unit 110 having two ends. A first end of the balancing unit 110 is connected to the test head 100 . The second end of the balance unit 110 is connected to the bracket 200 . Bracket 200 is a support structure for the test head around roll axis 102 .

If the center of gravity of the test head 100 does not coincide with the roll axis 102, then the test head 100 will rotate around the roll axis 102 away from the position shown in FIG. 2(a). The balance unit 110 will provide counter force to keep the center of gravity of the test head 100 balanced with respect to the roll axis 102 so as to maintain the position of the test head 100 in FIG. 2( a ). Thus, manual rotation of the test head 100 about the axis 102 is facilitated since the test head 100 is now substantially weightless relative to the roll axis 102 .

Figure 2(b) shows the same system as that shown in Figure 2(a), except that in Figure 2(b), bearings 116a and 116b are coupled together and connected to test head 100 at a single point outside. In contrast, in FIG. 2( a ), bearings 116 a and 116 b are connected to test head 100 , respectively.

Figure 3(a) shows another exemplary embodiment of the present invention. FIG. 3( a ) is a perspective view showing the application of the balancing unit 110 to the pivot shaft 102 of the test head 100 . The test head 100 is supported by a support arm 302 and a clamping flange 300 . Bearings 305 are accommodated inside the test head 100 . On the outside of the test head 100 , the support arm 302 is connected to the clamping flange 300 . There is provided a balancing unit 110 having two ends. A first end of the balancing unit 110 is connected to the test head 100 . The second end of the balancing unit 110 is connected to the clamping flange 300 . The clamping flange 300 is a support structure for the pivot shaft 102 .

If the center of gravity of the test head 100 does not coincide with the pivot axis 102, then the test head 100 will rotate about the pivot axis 102 passing through the bearing 305, away from the position shown in FIG. 3(a). The balance unit 110 will provide a reaction force to balance the center of gravity of the test head 100 relative to the pivot axis 102 and maintain the position of the test head 100 in FIG. 3( a ). Manual movement of the test head 100 is thus facilitated since the test head 100 is now substantially weightless relative to the pivot axis 102 .

Figure 3(b) shows the same system as that shown in Figure 3(a), except that in Figure 2(b), bearings 116a and 116b are coupled outside. In contrast, in FIG. 3( a ), bearings 116 a and 116 b are connected to test head 100 , respectively.

Although the invention has been described with respect to a single balancing unit 110 coupled to a single shaft 102, two or more units may be applied to a single shaft or to multiple shafts. When more than one unit is used, they can interact. The overall control scheme should preferably take this potential interaction into account. For example, if two units are used to control two independent pivot axes, they should be manipulated one at a time. It is contemplated that three or more units could be used successfully to control a test head mounted with ball bearings as described in the Background of the Invention.

Fig. 5(a) shows a substantially compliant drive mechanism 512 for a horizontal axis where there is little or no unbalanced force and the compliant force is mainly caused by friction, according to an exemplary embodiment of the present invention. In FIG. 5(a), flat panel 504 and flat panel 500 are representative of those used in, for example, PCT International Application No. US00/00704 to Holt et al., entitled "TEST HEAD MANIPULATOR", A plate in a structure with one or more built-in axes of horizontal motion for a manipulator base. As shown, two linear guides 503 a and 503 b are arranged parallel to each other and are rigidly attached to the upper surface of the plate 504 . Two pairs of ball sliders, a first pair including ball sliders 502a and 502b and a second pair including ball sliders 502c and 502d, are attached to the bottom surface of plate 500 . The first pair of ball sliders 502a and 502b are arranged such that they are slidably coupled to the linear guide 503a, and the second pair of ball sliders 502c and 502d are similarly arranged to be slidably coupled to the linear guide 503b. Thus, the plate 500 is parallel to and above the plate 504, and the plate 500 can move linearly relative to the plate 504 along the axis defined by the linear guides 503a and 503b. Such an arrangement allows for in-and-out motion or side-to-side motion in the test head manipulator base. It should also be noted that such an arrangement is also suitable for use with devices that rotate about a vertical axis as described, for example, in PCT International Application No. US00/00704 to Holt et al., entitled "TEST HEADNIPULATOR". Rocking motion caused by a horizontal plate relative to a horizontal plane. As shown in Figure 5(a), the resistance to motion is caused by the frictional forces associated with the ball sliders 502a, 502b, 502c, and 502d, and the linear guides 503a and 503b. For convenience, as shown in Figure 5 (a), the movement of the flat plate 50 along the linear guide rails 503a and 503b towards the direction of the linear actuator 508 is regarded as the movement in the "inward" direction, and the movement in the opposite direction The movement of is regarded as the movement in the "outward" direction.

A stationary element 506 including an actuator motor 510 , a linear actuator 508 is attached to the plate 504 . Typically, the linear actuator 508 includes a ball screw mechanism driven by an electric motor, and the driving component is a ball screw. However, other mechanisms and/or thread types may be used as appropriate for a given application. In FIG. 5(a), the driving component of the linear actuator 508 is the actuator shaft 514, which extends and/or retracts from the stationary element 506 when power is applied to the actuator motor 510. middle. The direction of motion of the actuator shaft 514 is determined by the direction of rotation of the actuator motor 510 . As shown, the linear actuator 508 is arranged to move the actuator shaft 514 in a plane parallel to the linear guides 503a and 503b.

The top end of the actuator shaft 514 has a small hole whose direction is nominally at right angles to the axial direction of the actuator shaft 514 . A clevis 516 and a clevis pin 518 passing through an aperture in the actuator shaft 514 couple the actuator shaft 514 and the compliant shaft 520 together. The axis of the compliant shaft 520 is preferably substantially coaxial with the axis of the actuator shaft 514; however, one of ordinary skill in the art will appreciate that other arrangements are possible.

The compliant shaft 520 passes through the lock 118 ; the lock 118 is rigidly attached to the plate 500 . The lock 118 may be one of many different types including electrically and pneumatically operated. When the lock 118 is engaged, the movement of the pad 500 relative to the compliance axis 520 , and thus relative to the pad 504 , is restricted. Thus, with the lock 118 engaged, the linear actuator 508 can be used to position the tablet 500 relative to the tablet 504 along the rails 503a and 503b. When the lock 118 is not engaged, the pad 500 is free to move relative to the compliance axis 520 and the pad 504; thus, compliant motion of the pad 500 relative to the pad 504 is achieved. The associated compliance forces due to this mechanism include the force required to overcome the friction of the linear guides 503a and 503b, and the ball sliders 502a, 502b, 502c, and 502d, plus the friction of the movement of the compliance shaft 520 through the locker 118. It should be noted that in a particular manipulator and application, some forces may additionally be present, however, these are not relevant to the mechanism described above.

Both extended compliant stops 522 and retracted compliant stops 524 are rigidly attached to the compliant shaft 520 , one on each side of the lock 118 . As shown in Figure 5(a), the extended compliant stop 522 is located between the lock 118 and the clevis 516; and the retracted compliant stop 524 is located between the distal end of the compliant shaft 520a and the lock 118 between. These compliance stops 522 and 524 act to limit the movement of the compliance axis 520 relative to the pad 500 , or, equivalently, limit the movement of the pad 500 relative to the compliance axis 520 . If the lock 118 is not engaged, the tablet 500 can be driven by an external force over a range of play defined by the distance "D" between the two compliant stops 522 and 524 minus the width "W" of the lock 118 movement, and this also limits the compliant range of movement. That is, with the lock 118 disengaged (and preferably closing the linear actuator 508), the panel 500 can be driven along the linear guide rails 503a and 503b over the distance defined by D-W under the external force sufficient to overcome the above-mentioned compliance force. sports.

In general, the friction associated with ball sliders 502a, 502b, 502c, and 502d, and linear guides 503a and 503b is greater than the friction between compliant shaft 520 and lock 118 . Thus, if the lock 118 is disengaged and the linear actuator motor 520 is energized, the compliant shaft 520 moves relative to the lock 118 and the plate 500 . If the actuator motor 510 is powered such that the actuator shaft 514 extends from the stationary element 506 , the extended compliant stop 522 moves in the direction of the lock 118 . If the actuator motor 510 is not closed, then, as the end of the compliance range is reached, the extended compliance stop 522 rests against the lock 118; then, further extension of the actuator shaft 514 causes the flat panel 500 to Plate 504 moves in an "out" direction. Additionally, if the actuator motor 510 is powered to retract the actuator shaft 514 into the stationary element 506 , the retracted compliant stop 524 moves in the direction of the lock 118 . If the actuator motor 510 is not turned off, then, as the end of the compliance range is reached, the retracted compliance stop 524 rests against the lock 118; then, further retraction of the actuator shaft 514 causes the tablet 500 to move relative to the Plate 504 moves in an "inward" direction.

During operation, compliance axis 520 is preferably positioned and locked in a position within a small neighborhood of a preselected point within its range of motion relative to tablet 500 . Typically, this point is midway between the two compliant stops 522 and 524; however, there are applications where it is more preferable to have this point located closer to either of the two compliant stops 522 or 524. Call this neighborhood the "compliant neutral region". The compliance neutral region is usually a small fraction of the overall compliance range; for example, ±3 or 4mm of the 40 to 50mm total compliance range.

When the lock 118 is located between the compliant neutral region and the extended compliant stop 522, it is generally said to be in the "compliant extension region". Similarly, when the lock 118 is located between the compliant neutral region and the retracted compliant stop 524, it is generally said to be in a "compliant retracted region".

As shown in FIG. 5( a ), a position sensor 528 is incorporated to detect the relative position of the plate 500 with respect to the compliance axis 520 . Position sensor 528 may be any of a number of known types ranging from precision encoders, potentiometers, etc., to combinations of limit switches, proximity sensors, range finders, and the like. As a minimum capability, the position sensor 528 should indicate which of three ranges the mechanism is currently in: compliant extension, compliant neutral, or compliant retract.

The operation of the compliance mechanism 512 is now described. First, the locker 118 is released. Linear actuator 508 is then used in conjunction with position sensor 528 to position compliant shaft 520 at a predetermined compliant neutral position. Now, the latch 118 is locked and the linear actuator 508 is used to position the tablet 500 in the desired position relative to the tablet 504 . For example, the desired location may be where a test head or load (not shown) will be docked. Now, the linear actuator motor 510 is turned off and the lock 118 is disengaged. Now, as the test head or load is driven into its final docked position, the plate 500 can move in a compliant manner driven by an external force, eg, a force generated by the docking mechanism. Depending on the particular application and/or user preference, the lock 118 may or may not re-lock once the final docked position is reached. Also, the lock 118 may or may not be locked while the test head or load is unlatched, depending on the particular application and/or user preference. However, once disengaged, the compliant shaft 520 should be placed and locked in the compliant neutral position according to the previous procedure before initiating motion of the pad 500 via the linear actuator 508 .

Figure 5(b) shows an alternative embodiment to Figure 5(a). In Figure (b), the hydraulic mechanism 513 replaces the locker 118 and the compliant stops 522 and 524 of Figure 5(a). In particular, the body of the double-acting hydraulic cylinder 140 is attached to the plate 500 and the axis of its cylinder is in a plane with the axis of the actuator shaft 514 of the linear actuator 508 . And, as shown in Figure 5 (b), the movement of the panel 50 along the linear guides 503a and 503b in the direction towards the linear actuator 508 is the movement in the "inward" direction, while the movement in the opposite direction is the movement in the direction Movement in the "outside" direction. The end of the cylinder 140 furthest from the linear actuator 508 is the “extended end” 532 , while the end of the cylinder 140 closest to the linear actuator 508 is the “retracted end” 533 . The piston 144 is located inside the hydraulic cylinder 140 ; and, the piston 144 is connected to the shaft 520 b which is coaxial with the cylinder barrel of the hydraulic cylinder 140 and protrudes through the retracted end 533 of the hydraulic cylinder 140 . Shaft 520b serves as part of compliant shaft 520 and it is connected to actuator shaft 514 via clevis 516 and clevis pin 518 as described with reference to Figure 5(a). In FIG. 5( b ), a second shaft 520 c attached to the opposite side of the piston 144 extends through the extended end 532 of the hydraulic cylinder 140 . This second axis 520c serves as an extension of the compliant axis 520, and the position sensor in FIG. 5(b) detects the position of this axis 520c relative to the plate 500. However, it is also feasible to arrange the system so that the position sensor 528 can operate in conjunction with the first shaft 520b to save the possible additional cost of providing the second shaft 520c.

Each of the two ends of the hydraulic cylinder 140 is provided with a port 154 to allow appropriate hydraulic fluid to flow into or out of the cylinder 140 . The two ports 154 are externally connected to each other by a conduit 152 via a valve 142 . Valve 142 is configured such that when it is open, fluid can freely flow out of one port and into the other port, and when it is closed, fluid cannot flow between the two ports. The entire mechanism is filled with a suitable, substantially incompressible fluid. The choice of fluid should take into account the possible deleterious effects that could occur if a leak were to develop in the system. Furthermore, since systems tend to be shipped in aircraft cargo holds at temperatures well below the freezing point of water, the fluid should have a low freezing point. Since fluid is incompressible, it is clear that when valve 142 is closed the piston cannot move in either direction. However, with valve 142 open, piston 144 is able to move because fluid is free to flow from one side of piston 144 to the other through the path provided by port 154, conduit 152, and open valve 142. When valve 142 is closed, the movement of compliant shaft 520 relative to hydraulic cylinder 140 and plate 500 to which cylinder 140 is attached is substantially locked. Thus, with valve 142 closed, operation of linear actuator 508 moves plate 500 relative to plate 504 along the axis defined by linear guides 503a and 503b. However, as the valve 142 is opened, the plate 500 can be driven by an external force to move within the range defined by the effective stroke of the piston 144 (substantially equal to the inner distance between the two end walls of the cylinder 140 minus the distance of the piston 144). thickness), moves along the axis defined by the linear guides 503a and 503b; and, this defines a compliant range of motion. That is, with valve 142 open (and preferably closed linear actuator motor 510), plate 500 can be driven by an external force along linear guides 503a and 503b over a distance defined by the effective stroke of piston 144. .

In general, with valve 142 open, the friction and other forces associated with movement of piston 144 within hydraulic cylinder 140 are less than the frictional forces associated with linear guides 503a and 503b, and ball sliders 502a, 502b, 502c, and 502d . Thus, if valve 142 is opened and linear actuator motor 510 is energized, compliant shaft 520 moves relative to hydraulic cylinder 140 and plate 500 . If the actuator motor 510 is powered to extend the actuator shaft 514 from the stationary member 506, the piston 144 moves toward the extended end 532 of the cylinder. Now, if the actuator motor 510 is not turned off, then, as the end of the compliance range is reached, the piston 144 bears against the end wall of the extended end 532 of the cylinder 140 . Further extension of the actuator shaft 514 then moves the pad 500 in an “outboard” direction relative to the pad 504 . If the actuator motor 510 is powered to retract the actuator shaft 514 into the stationary element 506 , the piston 144 moves toward the retracted end 533 of the cylinder 140 . Now, if the actuator motor 510 is not turned off, then, as the end of the compliance range is reached, the piston 144 bears against the end wall of the retracted end 533 of the cylinder 140 . Additionally, further retraction of the actuator shaft 514 moves the pad 500 in an “inward” direction relative to the pad 504 .

During operation, the piston 144 (and its associated compliant shaft 520) is preferably positioned and locked in a position within a small neighborhood of a preselected point within its range of motion relative to the plate 500. Typically, this point is midway between the two endpoints of the cylinder 140; however, there are applications where it is more preferable to have this point located closer to either of the two endpoints. We will again refer to this neighborhood as the "compliant neutral region". The compliance neutral region is usually a small fraction of the overall compliance range; for example, ±3 or 4mm of the 40 to 50mm total compliance range.

When the piston 144 is located between the compliant neutral region and the extended end 532 of the cylinder 140, it is generally considered to be in the "compliant extended region". Similarly, when the piston 144 is located between the compliant neutral region and the retracted end 533 of the cylinder 140, it is generally considered to be in a "compliant retracting region".

As in FIG. 5( a ), a position sensor 528 is incorporated to detect the relative position of the panel 500 with respect to the compliance axis 520 . Position sensor 528 may be any of a number of known types ranging from precision encoders, potentiometers, etc. to combinations of limit switches, proximity sensors, range finders, and the like. As a minimum capability, the position sensor 528 should indicate which of three ranges the mechanism is currently in: compliant extension, compliant neutral, or compliant retract.

The operation of mechanism 513 is now described. First, valve 142 is opened. Linear actuator 508 is then used in conjunction with position sensor 528 to position piston 144 in a predetermined compliant neutral position. Valve 142 is now closed, piston 144 and compliant shaft 520 are locked in the compliant neutral position, and linear actuator 508 is used to position plate 500 in the desired position relative to plate 504 . For example, the desired location may be where a test head or load (not shown) will be docked. Now, linear actuator 508 is closed and valve 142 is opened. Now, when the test head or load is driven into its final docked position, the plate 500 can move in a compliant manner under external forces, eg, forces generated by the docking mechanism. Depending on the particular application and/or user preference, valve 142 may or may not be closed once the final docked position is reached. Additionally, valve 142 may or may not be closed while uncoupling the test head or load, also depending on the particular application and/or user preference. However, once disengaged, the compliant shaft 520 should be placed and locked in the compliant neutral position according to the previous procedure before initiating motion of the pad 500 via the linear actuator 508 .

Figure 5(c) shows another exemplary compliant drive mechanism 550 for a horizontal axis where there is little or no unbalanced force and the compliant force is mainly due to friction. The arrangement and purpose of the two plates 500 and 504, linear guides 503a and 503b, ball sliders 502a, 502b, 502c and 502d, and linear actuator 508 are basically the same as described with reference to FIG. 5(a). Additionally, the compliant shaft 520 is connected to the actuator shaft 514 via the clevis 516 and the clevis pin 518 in a similar manner to that shown in FIG. 5( a ). As shown in Figure 5(c), the movement of the panel 50 along the linear guide rails 503a and 503b towards the direction of the linear actuator 508 is the movement in the "inward" direction, while the movement in the opposite direction is "outward". direction of movement.

The compliant shaft 520 passes through an inner mounting bracket 554 that is rigidly attached to the underside of the tablet 500 and an outer mounting bracket 552 that is closer to the linear actuator 508 than the outer mounting bracket 552 . Preferably, compliant shaft 520 is free to move along the in-out axis relative to mounting brackets 552 and 544, and ball sliders 502a, 502b, 502c, and 502d can be used to accomplish this task. A compliance stop 556 is rigidly attached to the compliance shaft 520 at a location between the two mounting brackets 552 and 554 .

The stationary element of the outer centering actuator 558 is attached to the outer mounting bracket 552 , while the stationary element of the inner centering actuator 560 is attached to the inner mounting bracket 554 . The driving part of each of the two centering actuators 558 and 560 is a plunger or ball screw that extends from or retracts into the stationary element once the actuator 558 or 560 is properly energized. Mechanism (for convenience, we use the term "plunger" to refer to each type of drive component where each is applicable and the ins and outs are clear). The two actuators 558 and 560 are mounted in opposite directions so that the plunger or ball screw is substantially coaxial and parallel to the axis of the compliant shaft 520, and so that when properly energized, the plunger or ball screw The mechanism protrudes toward the compliant stop 556 .

It can be seen that if one or both of the inner actuator plunger 564 and the outer actuator plunger 562 are retracted sufficiently so that the compliant stop 556 and one or both of the actuator plungers 562 and 564 If there is a gap between the distal ends of the orifice, then a compliant range of motion is formed. If this condition is satisfied, then the plate 500 can be made to overcome the friction of the ball sliders 502a, 502b, 502c and 502d, and the linear guides 503a and 503b in addition to the friction of the mounting brackets 552 and 554 and the compliant shaft 520. Driven by an external force, it moves along the linear guide rails 503a and 503b. The total range of compliant motion available is equal to the distance between the distal ends of the two actuator plungers 562 and 564 minus the thickness of the compliant stop 556 . Thus, in the system of Figure 5(c), the extent of compliance motion can be controlled by centering actuators 558 and 560, if desired; however, in the system of Figures 5(a) and 5(b), the compliance Regions are fixed.

It can also be seen that if both the inner and outer actuator plungers 564 and 562 are respectively extended so that both abut against the compliant stop 556 at the same time, then the plate 500 is fixed at a certain orientation relative to the compliant axis 520. position. Linear actuators 508 may then be used to move and position tablet 500 relative to tablet 504 .

In general, the friction associated with ball sliders 502a, 502b, 502c, and 502d, and linear guides 503a and 503b is greater than the friction between compliant shaft 520 and inner and outer mounting brackets 554 and 552, respectively. Thus, if one or both of centering actuator plungers 562 or 564 is retracted sufficiently to enable compliant motion and then power linear actuator motor 520, then the compliant shaft 520 and compliant stop 556 Movement is made relative to the inner and outer mounting brackets 554 and 552, respectively. If the actuator motor 510 is powered such that the actuator shaft 514 extends from the stationary element 506 , the compliant stop 526 moves in the direction of the outer actuator plunger 562 . If the actuator motor 510 is not turned off, the compliance stop 556 rests on the outer actuator plunger 562 as the end of the compliance range is reached. Further extension of the actuator shaft 514 then moves the pad 500 in an outward direction relative to the pad 504 . Conversely, if actuator motor 510 is powered to retract actuator shaft 514 into stationary element 506 , compliant stop 556 moves in the direction of inner actuator plunger 564 . If the actuator motor 510 is not turned off, the compliance stop 556 rests on the inner actuator plunger 564 as the end of the compliance range is reached. Further retraction of the actuator shaft 514 then moves the pad 500 relative to the pad 504 .

During operation, compliance axis 520 is preferably positioned and fixed at a position within a small neighborhood of a preselected point within its range of motion relative to panel 500 . Usually, this point is the midpoint of the two inner and outer centering actuators 560 and 558; . We will again refer to this neighborhood as the "compliant neutral region". The compliance neutral region is usually a small fraction of the overall compliance range; for example, ±3 or 4mm of the 40 to 50mm total compliance range.

When the compliant stop 556 is located between the compliant neutral region and the distal end of the outer actuator plunger 562, it is generally said to be in the "compliant outer region". Similarly, when the compliant stop 556 is located between the compliant neutral region and the distal end of the inner and outer actuator plungers 562, it is generally said to be in the "compliant inner region".

As discussed below, a useful system of the type in FIG. 5(c) can also be implemented without sensors to detect the relative position of plate 150 relative to plate 504 if proper centering actuator and design procedures are followed. However, a position sensor 528 (not shown) may optionally be included for this purpose, enabling a potentially more complex system. Position sensor 528 may be any of a number of known types ranging from precision encoders, potentiometers, etc., to combinations of limit switches, proximity sensors, range finders, and the like. As a minimum capability, the position sensor 528 should indicate which of three ranges the mechanism is currently in: compliant inside, compliant neutral, or compliant outside.

In order to implement a system without position sensor 528, it is preferable to use centering actuators 558 and 560 of this type, which can be controlled to move their plungers 562 and 564 along the given Move in a given direction until it reaches its full range of motion. Such actuators 558 and 560 could be as simple as a solenoid device with plungers 562 and 654 that, once energized, extend to their full extent, and once de-energized, under spring actuation, fully retracted. Actuators 558 and 560 comprising motors and screw mechanisms are also feasible. If hard stops are included in the actuators 558 and 560 to prevent overrunning of the screws, and if the motor is properly prevented from overheating when stalled, it can , simply drive the screw in a known direction. Additionally, the motor driven screw actuator may incorporate limit switches to signal that the end of the travel range has been reached. Another alternative could incorporate stepper motors and, by controlling the number of motor steps to control the position, possibly in combination with one or more limit switches. The second requirement is to design the stroke length of each actuator plunger 562 and 564, or screw, and the layout of the inner and outer mounting brackets 552 and 554 and centering actuators 558 and 560 to achieve the desired results. Specific criteria include: when both actuators 558 and 560 are in a fully extended state, the compliance stop 556 is firmly maintained in the desired compliance neutral position; and when the actuators 558 and 560 When both are in the fully retracted state, a supple range of motion is achieved.

The operation of the mechanism of Figure 5(c) without a position sensor is now described. First, centering actuators 558 and 560 are powered to fully extend their plungers 562 and 564 or screws. This moves the pad 500 relative to the compliance axis 520 to a compliant neutral position. Note that in this step of achieving the compliant neutral position in the system of Figure 5(c), the plate 500 moves while the compliant axis 520 remains stationary; however, in the systems of Figures 5(a) and 5(b) , the plate 500 remains stationary while the compliant axis 520 moves to achieve this. The actuators 558 and 560 are now in a state where their plungers 562 and 564 or screws remain in the fully extended position, causing the plate 500 to remain in the compliant neutral position. For motor-driven actuators that are not back-driveable, this can be accomplished by simply turning off the power. For other actuators, such as solenoid actuators, it may be necessary to maintain electrical excitation applied to them. With pad 500 thus held in the compliant neutral position, linear actuator 508 is then used to position pad 500 in a desired position relative to pad 504 . For example, the desired location may be where a test head or load (not shown) will be docked. Now, turn off the linear actuator motor 510 and control the centering actuators 558 and 560 in a manner that retracts their plungers 562 and 564 or screws to create a zone for compliant motion. Now, when the test head or load is driven into its final docked position, the plate 500 can move in a compliant manner under external forces, eg, forces generated by the docking mechanism. Depending on the particular application and/or user preference, centering actuators 558 and 560 may or may not be used to effectively lock tablet 500 relative to tablet 504 when the final docked position is reached. Furthermore, the tablet 500 may or may not be held in a position by the centering actuators 558 and 560 while the test head or load is disengaged, also depending on the particular application and/or user preference. However, once disengaged, the compliant shaft 520 should be placed and locked in the compliant neutral position according to the previous procedure before initiating motion of the pad 500 via the linear actuator 508 .

A system according to Fig. 5(c) also comprising a position sensor 528a (not shown) for detecting the position of the panel 500 relative to the compliance axis 520 is now described. In such a system, it is most practical to use electric motors to operate the centering actuators 558 and 560 in order to take advantage of the effective feedback position. The overall travel of the actuator plungers 562 and 564 or screws and the placement of the actuators 558 and 560 are not as important as in a system without the position sensor 528a as described above. Actuators 558 and 560 are arranged to: when both of their plungers 562 and 564 are fully retracted, achieve a suitable compliant range of motion; and when they are fully extended, plungers 562 and 564 reach at least a compliant neutral position Will suffice. Operation is essentially the same as the system without the position sensor 528a, except that the position sensor 528a is used to indicate when the compliant stop 556 has reached the compliant neutral position. Furthermore, if an encoder or potentiometer type position sensor 528a is used, in the system of Fig. 5(c), as well as in the systems of Fig. 5(a) and Fig. 5(b), by suitably programming the control system, no change The mechanical structure can change the position of the soft and neutral area.

The mechanism of Figure 5(a) (and Figures 5(b) and 5(c)) can also be applied to pivot axes, where the motion is rotation of the body or load about the axis, rather than translation along the axis. This technique can also be applied to a combination of rotation and translation. FIG. 7 shows the compliance mechanism applied to a body 100 rotating about a vertical axis 102 . This is the same as the deflection motion in the test head manipulator. The compliant mechanism in Figure 7 also shows an alternative embodiment to the mechanism in Figure 5(a). Specifically, the mechanism in Figure 5(a) has an actuator shaft 514 or screw attached to an arm (compliant axis 520) and a movable body attached to a lock 118 operating on the arm (compliant axis 520). (Tablet 500). In Figure 7, the arrangement of the locker 118 and the arm (rod 112) is reversed; Coupled with the movable body 100. It can be seen that the operation in Fig. 7 is completely similar to that in Fig. 5(a). The features included in Figure 7 are discussed in more detail below.

The embodiment of Figures 5(a) to 5(c) has set forth an exemplary embodiment illustrating compliance to the axis of motion in the horizontal plane or otherwise in the absence of any significant unbalanced forces. Figures 6(a) and 8(a) illustrate the ability to provide balanced forces that counteract unbalanced forces, and compliance in drive shafts where unbalanced forces can be significant, vary with position, and often unpredictable. mechanism. These mechanisms are known as "Balanced Compliant Actuator Assemblies (BCAAs)".

In Figures 6(a) and 8(a), the body 100 represents a heavy load (i.e., a test head) that is rotatable about a pivot axis 102 located near a surface at a considerable distance from the body's center of gravity . The main body 600 in FIGS. 6( a ) and 8 ( a ) represents a portion of a fixed structure supporting the pivot shaft 102 . For example, body 100 may represent a test head with a pitch or roll axis near the dut interface, while body 600 may represent a portion of a test head mount that is in turn attached to the main arm of a test head manipulator (not shown). Connected equipment supports. In the following discussion, it is assumed that the pitch axis 102 is in the horizontal plane unless otherwise stated. Note that if the body 100 is rotated such that its center of gravity is not directly below the pitch axis 102, then the moment caused by the weight of the body 100 causes an imbalance that varies with the position of the body 100. Thus, the compliance force includes the force required to overcome this moment as well as friction and other possible effects such as rope forces.

The mechanism shown in Figures 6(a) and 8(a) and to be described is included in the title "A Test Head Balancing System for a Test Head Manipulator (A Test Head Balancing System for a Test Head Manipulator), awarded to Ny et al. people and assigned to the same assignee as the present application, the U.S. Provisional Patent Application No. 60/234,598 illustrates the concept. Specifically, in the Provisional Application No. 60/234,598, Fig. 1(a), 1( c), 2(a), 3(a), 4(a) and 4(b).

Although not shown, it is implied in Figures 6(a) and 8(a) that a system controller is present. The system controller executes the control sequences and provides control signals to the actuators and valves, as will be described later. In addition, the system controller receives operator input and feedback signals from sensors, which are also described below.

In summary, both Figures 6(a) and 8(a) include a linear actuator 508 that drives the body 100 about the pivot axis 102, including the lock 118, the lever 112, and two compliant stops 522 and 524. A compliant mechanism for compliant motion of the main body 100, a force sensor 120 for measuring unbalanced forces applied in the force bar 112, and a double acting pneumatic cylinder 128 for generating a balanced force to compensate for the unbalanced forces.

Figures 6(a) and 8(a) are quite similar. The only difference between them is how the body of the double-acting pneumatic cylinder 128 is supported. Therefore, the common features of these two figures are described together below; and the differences between them are described separately when appropriate.

In both FIGS. 6( a ) and 8 ( a ), the stationary element 506 of the linear actuator 508 , including its motor 510 , is attached to the main body 600 by means of bearings 620 . The driving part of the linear actuator 508 is the actuator screw 514, which extends from or retracts into the stationary element 506 when power is applied to the motor 510; 510 energizes the motor to rotate in a first direction of rotation and the screw (actuator) protrudes from the stationary element 506; and if power is supplied to the motor 510 to rotate the motor in the opposite direction, the screw (actuator) Retracts into stationary element 506 . Actuator screw 514 is typically a ball screw; however, other types of screws or other actuator devices may be used where appropriate.

The force rod 112 is attached to the main body 100 by means of a bearing 116a. As shown, the force rod 112 is engaged by a lock 118 which is rigidly attached to the distal end of the actuator screw 514 by means of a machine screw or other suitable means. The lock 118 may be any of several types well known in the art. Depending on the type of interlock selected, it may be controlled by electrical signal, gas input, or other means appropriate to the particular application. When the lock 118 is actuated, it firmly grasps the force rod 112 and then effectively attaches the force rod 112 to the actuator screw 514 in a rigid manner. Accordingly, movement of the body 100 relative to the actuator screw 514 is restricted when the locker is engaged. When the lock 118 is released, the force rod 112 can slide along the lock 118 and move along a line substantially parallel to the axis of the actuator screw 514 . Thus, the linear actuator 508 can be used to rotate the main body 100 about the pivot axis 102 relative to the main body 600 by actuating the lock 118 while momentarily setting aside the pneumatic cylinder 128 and any unbalanced forces. Additionally, when the lock 118 is not locked, the body 100 responds to sufficient external force to move freely relative to the actuator screw 514 and the body 600; thus, compliant movement of the body 100 relative to the body 500 is achieved. That is, as the lock 118 is released, compliant movement of the body 100 about its pivot axis 102 and relative to the body 600 is achieved in response to an external force. To limit the range of compliance motion, an extended compliance stop 522 and a retracted compliance stop 524 are rigidly attached to the force rod 112 . A protruding compliant stop 522 is located between the lock 118 and the main body 100 . The retracted compliant stop 524 is located opposite the lock 118 and between the lock 118 and the distal end of the force rod 112 .

Bi-directional force sensor 120 is coupled to force bar 112 in such a way as to measure force along force bar 112 . Force sensor 120 may be implemented with a bi-directional load cell, a readily available device. The load cell comprises strain gauges which may be placed in a bridge circuit in a well known manner, providing a voltage output which varies monotonically with the measured force. Using an analog-to-digital converter and a processor, it can be determined whether the force along the force bar 112 is greater than the maximum tolerable for free motion; and, if so, the direction of the force. Alternatively, an analog comparator circuit may be used in known manner to generate a yes/no signal indicative of the presence and direction of a significant unbalanced force.

The aforementioned compliant stops 522 and 524 are used to limit the movement of the force rod 112 relative to the main body 600 when the linear actuator 508 is extended to a certain length. If the rod lock 118 is not engaged, the force rod 112 can be driven by an external force acting on the body 100 to pass the distance "C" between the two compliant stops 522 and 524 minus the width of the rod lock 118 "". L" defines the range of play, and this also defines the range of motion compliance of the main body 100. That is, with the lever lock 118 released (and preferably turning off the linear actuator motor 510), the main body 100 can be driven by the force lever 112 about the pivot axis 102 under the drive of an external force sufficient to overcome the compliance force. The angle determined by the corresponding movement over the distance C-L. In Figure 6(a) or 8(a), as the lock 118 moves relative to the force rod 112, the piston 130 in the pneumatic cylinder 128 will also move. It is important that the total effective stroke of piston 130 be designed such that piston 130 and pneumatic cylinder 128 do not interfere with compliant stops 522 and 524 .

It can be seen that if the lock 118 is disengaged and the main body 100 is turned at an angle so that its center of gravity is not just below the pivot axis 102, then one of the two compliant stops 522 and 524 will rest against the lock. device 118. It is a matter of whether the moment about the pivot axis 102 caused by the weight of the body 100 is sufficient to overcome the static friction in the system, the forces acting on the body 100 due to the pneumatic cylinder 128 and piston 130, and any forces acting on the system. In this state, it will be apparent that energizing the actuator motor 510 will cause the main body 100 to rotate about the pivot axis 102 . This is not necessarily the best mode of operation; rather, it is something to be aware of when considering various sequences of operations for exemplary embodiments of the present invention.

During operation, the lock 118 is preferably moved and locked into position within a small neighborhood of a preselected point within its range of motion relative to the lever 112 . Typically, this point is the midpoint of the two compliant stops 522 and 524; however, there are applications where it is more preferable to have this point located closer to either of the two compliant stops 522 or 524. We will refer to this neighborhood as the "compliant neutral region". The compliance neutral region is usually a small fraction of the overall compliance range; for example, ±3 or 4mm of the 40 to 50mm total compliance range.

Still referring to Figures 6(a) and 8(a), when the lock 118 is located between the compliant neutral region and the extended compliant stop 522, it is generally said to be in the "compliant extension region". Similarly, when the lock 118 is located between the compliant neutral region and the retracted compliant stop 524, it is generally said to be in a "compliant retracted region".

It may be necessary to balance the body 100 in order to disengage the lock and then move to the compliant neutral position. This balancing can be achieved in many possible ways including, for example, having the center of gravity below the pivot axis 102 and/or applying an external force. However, as explained above, the purpose of the pneumatic cylinder 128 and piston 130 is to provide a controllable source of balanced force.

A position sensor 612 is provided to detect the position of the force bar 112 relative to the lock 118 . Position sensor 612 may be any of a number of known types ranging from precision encoders, potentiometers, etc., to combinations of limit switches, proximity sensors, range finders, and the like. As a minimum capability, the position sensor 612 should indicate which of three ranges the mechanism is currently in: compliant extension, compliant neutral, or compliant retract.

A double-acting pneumatic cylinder 128 is provided to generate force against any unbalanced forces. Its purpose is to minimize the net forces and/or moments acting on the body 100 to acceptable levels in order to allow compliant movement of the body 100 relative to the body 600 . When this is achieved, the body 100 may be considered to be in a "state of equilibrium". Note that when in equilibrium, there may be "residual imbalance forces" that become components of the total compliance force. In FIG. 6( a ), double acting pneumatic cylinder 128 is attached to main body 600 by means of suitable mounting hardware 624 and bearings 622 . In FIG. 8( a ), a double-acting pneumatic cylinder 128 is attached to the locker 118 using a bracket. In FIGS. 6( a ) and 8 ( a ), the pneumatic cylinder 128 includes a piston 130 coupled to a connecting rod 114 that is coaxial with the axis of the cylinder 128 . The connecting rod 114 extends toward the main body 100 through the end of the pneumatic cylinder 128, and it is coupled with the main body 100 through the bearing 116b. Although Figures 6(a) and 8(a) show the connecting rod 114 and force bar 112 attached to the main body 100 by means of two separate bearings 116a and 116b, as otherwise identical to Figure 8(a) As shown in Figure 8(b), other alternative configurations are possible including the use of brackets or the like to co-couple both rods 114 and 112 with the body 100 by means of bearings 116a and 116b. FIG. 8( b ) shows locker 118 and cylinder 128 coupled together with bracket 800 , which effectively couples rods 114 and 112 .

A high pressure gas supply 602 provides a gas source for operating the pneumatic cylinder 128 . Typically, an air pressure of 80 to 100 pounds per inch is suitable and achievable in a semiconductor test facility. The pneumatic cylinder 128 contains two gas inlets 132 and 134 , one on either side of the piston 130 , generally near each end of the cylinder 128 . The first gas inlet 132 is coupled to the first accumulator 608a by means of a conduit, and the second gas inlet 134 is coupled to the second accumulator 608b by means of a conduit. The first and second accumulators 608a and 608b are coupled to the first and second valves 604a and 604b, respectively, by means of conduits, while the first and second valves 604a and 604b are coupled to the gas supply 602 by means of appropriate conduits. The first and second valves 604a and 604b are controlled by a controller not shown.

In addition, the two valves 604a and 604b each have a vent to vent gas out of the cylinder 128 and accumulators 608a and 608b. Each of valves 604a and 604b can be in one of three positions:

The "injection position" allows gas to flow from gas supply 602 into accumulators 608a and 608b and cylinder 128 .

The "vent position" vents gas from accumulators 608a and 608b and cylinder 128 to atmosphere through the vent ports.

The "closed position" prevents gas from flowing into or out of accumulators 608a and 608b and cylinder 128 through valves 604a and 604b.

Many different configurations of valves (604a and 604b) are commercially available. For example, instead of a single separate valve (604a and 604b) as shown, there could be a unit containing two valves configured such that if one valve is in the fill position and the other is in the discharge position, Then, when gas is injected from one side of the piston 130, gas is automatically discharged from the opposite side.

As noted above, the purpose of the pneumatic cylinder 128 is to generate a force that compensates for unbalanced forces acting on a load, which in this example is the body 100 . For example, assume that the main body 100 is pivoted at an angle such that its center of gravity deviates from the pivot axis 102 in the horizontal direction. The weight of the body 100 is then an unbalanced force that creates a torque about the pivot axis 102 acting in a direction that moves the center of gravity to a point directly below the pivot axis 102 . The magnitude and direction of this torque is a function of the weight, and the magnitude and direction of the angle of rotation, where the angle of rotation is the inverse of the ratio of the center of gravity relative to the horizontal position of the pivot axis 102 for the shortest distance from the center of gravity to the pivot axis 102 string. The unbalanced force is seen to be a variable and a function of the load position. The pneumatic cylinder 128 generates a force on its piston 130 based on the difference in air pressure on either side of the piston 130 . This force is transmitted along the link 114 acting on the body 100 at a position and orientation determined by the extension of the axis of the continuous rod 11 . The device is arranged such that this line of force does not intersect the pivot axis 102 throughout the range of rotation angles of interest. Therefore, the piston 130 can generate a torque on the main body 100 about the pivot axis 102, which torque is equal to the unbalance torque caused by the weight of the main body 100, and the direction is opposite to the unbalance torque caused by the weight of the main body 100, thereby, Bring the main body 100 to a state of equilibrium. The forces generated by pneumatic devices are referred to herein as "balanced forces".

During operation, valves 604a and 604b are normally in a closed position. In order to increase the pressure acting on one side of the piston 130, the corresponding valve (604a or 604b) is switched to the injection position. The pressure change obtained is a function of the current pressure in the cylinder 128, the pressure of the gas supply, and the length of time the valve (604a or 604b) is operating. To reduce the pressure acting on a given side of the piston 130, the corresponding valve (604a or 604b) is switched to the exhaust position. Here, the pressure change achieved is simply a function of the current pressure in the cylinder 128 and the length of time the valve (604a or 604b) is operating. Thus the minimum length of time that the valve (604a or 604b) operates determines the minimum pressure increase in the cylinder 128 that can be achieved for a given initial cylinder pressure. Typically, the minimum time is 8 to 10 microseconds. The minimum counterbalance force increment that can act on the connecting rod 114 is determined by the product of the minimum increment of pressure and the area of the piston 130 . With proper selection of components and parameters, balance force increments of less than 2 lbs can be achieved, and this determines how as close as possible the unbalanced forces can be canceled out.

If a counterbalancing force is established by the pneumatic cylinder 128 and piston 130 with the body 100 in a particular position relative to the body 600, and if the body 100 is now moved slightly relative to the body 600, the piston 130 moves within the cylinder 128 , and the counteracting force that builds up will vary as the volume of gas on each side of the piston 130 changes. Specifically, the pressure differential increases monotonically with displacement of the piston 130, similar to the mechanism spring. The aerodynamic force exerted on a given side of the piston 130 varies inversely with volume. However, for displacements where the volume change of the gas is relatively small, the equivalent elastic force varies approximately linearly with displacement, as well as the equivalent "spring constant" K. That is, F=Kx, where F is the change in force and x is the change in piston 130 displacement. Preferably, try to keep K small so that the force does not vary considerably over the range of motion. K is determined in part by the change in pressure of the total volume of gas relative to each incremental displacement of the piston 130, that is, by V and dP/dx, where V is volume and P is pressure. As the piston 130 approaches one end of its stroke, V becomes small and dP/dx becomes exponentially larger, so K increases like F. An accumulator as described above can be added to significantly increase V and limit dV/dx and dP/dx to acceptable values. This provides a relatively low and constant value of K in the compliance range, or the so-called "soft spring" effect, which is clearly desired.

Having briefly described the components of the apparatus shown in Figures 6(a) and 8(a), the details and operation of the system are now further explained.

To achieve a balanced state, the locker 118 is first locked to prevent the main body 100 from moving. The controller then operates the valves 604a and 604b in an appropriate manner in response to the signal from the force sensor 120 so that a counterbalancing force is applied to the body 100 . The controller adjusts the balance force until it has sufficient magnitude and direction to reduce the magnitude of the force in the force rod 112 to a magnitude less than a predetermined threshold, which represents the minimum allowable unbalanced force. Depending on the environment, this threshold is usually less than 5 to 10 pounds; and, if the design of the system is considered, it can be as low as 1 to 2 pounds. When this condition has been satisfied, the body 100 is said to be in a "balanced state" and the difference between the original unbalanced force and the balanced force generated by the piston 130 will be referred to as the "residual unbalanced force".

Assume now that the body 100 is in such a balanced state with the lock 118 locked and positioned away from one of the two compliant stops 522 and 524 . If the lock 118 is now released, the main body 100 and lock 118 will not move provided any residual unbalanced forces are less than the static friction in the system, including the effect of breakaway forces associated with the pneumatic cylinder 128 and piston 130 . If this state is achieved, then the subject 100 is considered "perfectly balanced." However, if the residual imbalance force is sufficient to overcome all static friction forces, then the body 100 is considered "nearly balanced". If the body 100 is nearly balanced and the lock 118 is released, the lock 110 will move towards one of the two compliant stops 522 or 524 . Whether the lock 118 rests against the compliant stop (522 or 524) depends on how both the unbalanced and balanced forces change as a function of the change in position. It should be understood that achieving a fully balanced state requires greater accuracy, and thus, higher system cost, then, to achieve a near balanced state.

The term "compliance ready state" is now introduced to indicate the situation when the main body 100 is in equilibrium and the lever 112 is in the compliant neutral region. Generally speaking, the best equilibrium state is a perfect equilibrium state, so that if the lock is disengaged, the main body 100 does not move.

In general, it can be assumed that the main body 100 is not in a balanced state initially, but is at any position relative to the main body 600 . It is also assumed that a state of compliance readiness is to be achieved. In this case, the unbalanced force is supported and measured by force rod 112 and force sensor 120 . If the lock 118 is in the disengaged state and if the unbalanced force is substantial, then the lever 112 is most likely at this or that end of its compliant range; Locker 118 on (122 or 124) communicates from force rod 112 to actuator screw 514. However, if the lock 118 is in the locked state, then the unbalanced force is transferred directly from the force rod 112 to the lock 118 and thus to the actuator screw 514 .

There are several possible ways of achieving the desired compliant readiness. However, before any control action, care should be taken to ensure that the main body 100 and other equipment are not resting on any foreign objects or structures that may interfere with the upcoming balancing and compliant range positioning operations. In some cases, it may be necessary to first move the main body 100 into some desired position. In such a case, if the lock 118 is not already actuated, the lock 118 is first actuated and then the linear actuator 508 is used to generate the desired movement. Movement of other manipulator axes, if present, may also be necessary. The movement may be accomplished by manual means, which may include commanding the controller to manipulate the linear actuator 508 using button commands. Optionally, the controller may be equipped with appropriate algorithms enabling it to utilize the linear actuator 508 in combination with position feedback in an automated positioning sequence. With the main body 100 in the desired position, there are several possibilities for reaching a compliant readiness. The design and choice of a specific method will depend on the details of the existing application. Described below are the two selected methods, along with notes indicating the possible application of each. In both cases, it is assumed that the lock 118 is not in the compliant neutral position.

Method 1: As shown in the flowchart in Figure 15(b), this method needs to design the system so that the balance result always leads to a completely balanced state, which includes the following series of steps:

1. As shown in step 1510, if the locker 118 is not already locked, then lock it.

2. As shown in step 1511, the main body 100 is brought into equilibrium according to the aforementioned process.

3. As shown in step 1512, the lock 118 is unlocked.

4. As shown in step 1513, the linear actuator 508 is used to position the lock 118 in the compliant neutral position.

5. As shown in step 1514, lock the locker 118 again.

6. As shown in step 1515 , bring the body 100 into balance again to adjust for any changes in imbalance caused by step 1513 .

7. As shown in step 1516, the main body 100 is now fully balanced and in a compliant ready state with the lock 118 locked.

Note that step 1513 needs to meet two conditions to be realized. First, it is required that the equilibrium achieved in step 1513 be a state of perfect equilibrium so that when the lock 118 is released, the main body 100 remains stationary. Second, the friction between the locker 118 and the force rod 112 is required to be less than the static friction of the rest of the system, including the breakaway force of the piston 130 in the pneumatic cylinder 128 .

Approach 2: As shown in the flowchart in Figure 15(c), this approach does not require designing the system such that equilibrium always leads to a perfectly balanced state. It can work when the balance is perfectly balanced or nearly balanced. It includes the following series of steps:

As shown in step 1520, the latch 118 is unlocked.

As shown in step 1521, based on feedback from position sensor 612, controller operates valves 604a and 604b to manipulate the gas pressure in double-acting pneumatic cylinder 128 to move piston 130 along the direction required to move force rod 112 relative to lock 118. direction to move. This causes the force bar 112 to move relative to the lock 118 in the desired direction.

As shown in step 1522, when the compliant neutral position is reached, the movement in step 1521 is aborted. The lock 118 may act as a brake to stop movement.

As shown in step 1523, the locker 118 is locked.

As shown in step 1524, the main body 100 is brought into equilibrium according to the process previously described.

As shown in step 1525, the main body 100 is now in a compliant state, but not necessarily in perfect balance.

Note that in steps 1521 and 1522, the piston 130 is the initial mover. The control algorithm for positioning the force rod 112 relative to the locker 118 can be difficult to implement in terms of system stiction including inter alia the breakaway force of the piston 130 (note that in method 1 the initial pusher is controlled in the linear actuator 508, and , which usually provides enough force to overcome the system static resistance). This is especially true if an attempt is made to stop the main body 100 in the desired position prior to application of the lock 118 . However, suitable locks 118 can also be used as brakes, and as the body 100 is moved to the desired compliant neutral position, the locks 118 can be applied to force the movement to stop.

In a test head manipulator system, one of two possible actions is likely after reaching the compliant ready position. A first possibility is that the test head 100 is ready for docking. In this case, it can be moved slightly so that the docking actuator and docking guide mechanism (not shown in Figures 6(a) and 8(a)) engage first. Docking devices are typically designed in such a position that the delicate electrical contacts are not yet engaged and there is no risk of damage. This small amount of movement is usually not enough to interfere with the compliant ready position. If the latch 118 is locked, it can now be released. Movement only results in contact being made between cooperating guide parts of the docking device if there is a slight residual unbalanced force sufficient to cause movement. The main body 110 is now able to move compliantly, and the docking actuators can now engage, drawing the test head 100 into the fully docked position.

A second possibility is to move the test head 100 from its current position to a new position by means of a manipulator device (not shown in Figures 6(a) and 8(a)). In this case, the latch 118 is preferably locked to maintain the compliant neutral position, and the linear actuator 508 can be used as needed to move the body 100 to the desired second position. It should be noted that as the body 100 is thus moved, the unbalanced forces acting on it will likely change. Then, when the main body 100 reaches said second position, it should be brought into equilibrium again. Once balanced, the main body 100 is in a compliance-ready state since it is still locked in the compliance-neutral position. If desired, the lock 118 can be released and the main body 100 can move compliantly about the pivot axis 102 .

Ideally, the process of reaching a compliant readiness state is done only once and not repeated until some change is made to the system including, for example, the position of body 100 relative to body 600 . However, pneumatic cylinder systems are prone to gas leaks and corresponding pressure loss, so the process required to achieve equilibrium or compliance readiness should be repeated periodically when required. In an exemplary embodiment, the system maintains sufficient gas for up to about 10 minutes or more before the desired process needs to be repeated. In typical operation, the cycle repeats automatically every few minutes (say, 5 to 8 minutes). Another consideration caused by gas leaks and loss of gas differential pressure is that when the test head is docked for an extended period of time, the balance force will be lost. This can happen, for example, when tests last many hours or many minutes. When docked, the test head is held securely by the docking mechanism, and the balancing forces cannot be re-established by the aforementioned methods. Therefore, for disengagement, the locker 118 must first be locked in order to securely hold the body 100 relative to the actuator screw 514 . Then, the disengagement movement driven by the docking actuation mechanism should be such that no relative movement between the bodies 600 and 100 is required. As an alternative, the system can be equipped with optional pressure sensors (600a, 606b) and, as docking begins, the pressure of the cylinder 128 can be recorded. Then, just before disengagement, cylinder 128 pressure can be restored so that proper balance can be restored. This releases the lock 118 and allows the body 100 to move compliantly while disengaging.

So far, the pivot axis 102 has been discussed in the horizontal plane, which is sufficient to illustrate the principles of the present invention. However, it is also possible that the pivot axis 102 may be in other orientations. The specific structure considered is the case of a vertical pivot axis, where the body 100 represents the support structure for the manipulator cylinder, which in turn supports the test head. In this case, the pivot axis 102 corresponds to the deflection axis of the manipulator. Unbalanced forces may exist in such shafts due to the forces generated by the bending of the test head cord. In contrast to the aforementioned unbalanced forces due to gravity, in the absence of gravity in this case, the unbalanced forces induced by the ropes will vary in a generally unpredictable manner. Nonetheless, it is clear that the mechanism of Figure 6(a) or Figure 8(a) can still be used effectively to provide a drive deflection axis that is compliant and counterbalanced by the force of the cable. For example, Figure 6(b) shows a BCAA of the type shown in Figure 6(a) applied to a pivot axis, in this case the pivot axis being vertical, ie the deflection axis. It can also be seen that if the rope forces are negligibly small, then a simplified version of the mechanism which does not include the double acting pneumatic cylinder 128 and associated equipment shown in Figure 6(a) can be effectively used. Fig. 7 thus shows the application of a deflection axis 102 which does not require any counter force. The difference between Fig. 7 and Fig. 6(a) is that the force sensor 120 and the counter force device of the pneumatic cylinder 128 are removed. FIG. 7 is similar to the compliance mechanism of FIG. 5( a ), except that in FIG. 7 the actuator rod 514 is coupled directly to the lock 118 rather than to a rod such as rod 520 .

As mentioned above, the BCAAs of Figures 6(a) and 8(a) are quite similar. Which of these two structures to use is usually based on the details of the physical arrangement of the purpose. For example, Figure 8(a) would be suitable for situations where the desired range of motion in the shaft 102, and thus the linear actuator 508, is much greater than that of compliant motion. This is because, in FIG. 8( a ), the stroke of piston 130 within cylinder 128 need only be sufficient to accommodate compliant motion; Thus, a smaller and less expensive pneumatic cylinder 128 and piston 130 assembly can be used in Figure 8(a). Figure 6(a) is the preferred example where there are space constraints that would prevent all the components of the BCAA from being placed close together. In particular, in Figure 6(a), the pneumatic cylinder 128 and piston 130 subassembly may be located remotely from the subassembly including the actuator 508, lock 118, force rod 112, and associated components and sensors.

Figure 9(a) shows an example of a typical application of two Balanced Compliant Actuator Assembly (BCAA) mechanisms as shown in Figure 8(a) used in a typical manipulator system application. In FIG. 9( a ), one mechanism affects the yaw axis 934 of the test head 100 while a second mechanism affects the pitch axis 936 of the same test head 100 . A single control system, not shown for simplicity, is also provided to achieve the desired control of the two BCAAs as well as the entire manipulator system.

In FIG. 9( a ) a test head 100 is shown having a test interface on the top surface 901 . The DUT board 900 provides the interface circuitry and connection equipment needed to properly interface the device under test with the test head pin electronics (not shown) located inside the test head 100 .

Docking assembly 916 including three of four docking cams 910 , three of four guide pins 912 , rope 915 , rope driver 917 , and a portion of docking controller 914 are also shown in cutaway. In general, to interface test head 100 with test equipment, such as a device handler (not shown), wafer inspector, etc., test head 100 is first manipulated so that guide pins 912 are partially inserted into mating docking assemblies attached to the test equipment. (not shown) in the mating guide hole (not shown). Here, the docking cam 910 engages a mating cam follower (not shown) on a mating docking assembly (not shown). Now, an external force can be applied to rotate the docking handle 914, thereby rotating the cord driver 917. As the cable drive 917 rotates, its motion is transmitted through the cable 915, causing the docking cam 910 to rotate synchronously, thus pulling the test head 100 into its final docking position. While docking handler 914, cable drive 910, and docking cam 910 are pulling test head 100 into position, it is desirable for test head 100 to move compliantly in all six degrees of freedom. This docking is shown in more detail in Figures 12(a) to 12(d) to be discussed later. Furthermore, this docking is a modification of the docking described in US Patent No. 4,489,815 to Smith, which is hereby incorporated by reference, using two guide pins 912 and cams 910 . Also, other docking members of this type use three guide pins 912 and cams 910 . See the inTEST manual for further information.

The test head 100 is mounted on a typical cable pivot ring 924 of a test head manipulator (not shown) as described, for example, in US Patent Nos. 5,030,869 and 5,450,766 to Holt and the inTEST manual. FIG. 9( a ) shows a test head mount 926 attached to a rope pivot ring 924 to which a first end of a test head support beam 922 is attached. The axis of the test head support beam 922 is normal to the plane of the cable pivot ring 924 and the extension of the axis passes through the center of the cable pivot ring 924 . Test head support beam 922 extends through aperture 920 on the rear of test head 100 and is rigidly attached to outer ring 904 of circular bearing assembly 905 . An inner ring 902 is arranged inside and is supported by an outer ring 904 of a circular bearing assembly 905 . Inner and outer rings 902 and 904 are circular, coplanar, and concentric, respectively. Roll axis 934 is perpendicular to the plane of the two rings 902 and 904 and passes through their common center. The inner ring 902 rotates about the yaw axis 934 inside the outer ring 904 . Two accessory shafts 906 pass through gaps in the outer ring 904, coupling the inner ring 904 with bearing blocks 908, which are secured on either side of the test head 100, which are shown at perpendicular angles parallel to the test head support beam 922. on flat surface. Bearing block 908 enables test head 100 to rotate about pivot axis 936 defined by accessory shaft 906 .

Six cam followers (not shown) are used as support means to take up the load exerted by the inner ring 902 on the outer ring 904 while performing a low friction yaw motion. Three of the cam followers (not shown) are located 120° apart near the inner circumference of the outer ring 904 and bear radial loads in a direction perpendicular to the yaw axis 934 . The other three cam followers (not shown) are also arranged inside the outer ring 904 at equal intervals between the first three in a manner to support an axial load parallel to the yaw axis 934 .

Overall, the tether pivot ring 924 supports the test head mount 926 . A test head mount 926 supports a test head support beam 922 that defines a roll axis. Test head support beam 922 supports outer ring 904 which in turn supports inner ring 902 . Inner ring 902 provides yaw axis 934 and supports accessory shaft 906 . Auxiliary shaft 906 supports test head 100 and provides pitch axis 936 . This arrangement enables test head 100 to rotate ±5° about pitch axis 936 (shown horizontal) and ±5° about roll axis 934 (shown vertical), where the yaw and pitch axes are relative to each other. Perpendicular to the test head support beams 922 and to each other. The ranges of the pitch and yaw axes are limited to within ±5°, respectively, by physical constraints and mechanical stops built into the system not shown in Figure 9(a) for simplicity. Additionally, position sensors (not shown) may be included in known manner to indicate the relative position of test head 100 with respect to pitch axis 936 and roll axis 934, and to provide this information to the control system.

The tether pivot loop 924 generally provides ±95° roll of the test head 100 about the axis defined by the support beam 922 . In Figure 9(a) with the test interface surface facing upwards, the test head 100 is shown at approximately +90°. Components of the test head cord are threaded through cord pivot ring 924 and opening 928 in test head mount 926 is arranged along test head support beam 922 and into test head 100 through support rod opening 902 . Webs 918 attached to support beams 922 provide support and strain relief for the tether elements as they disperse to various electronic components (not shown) inside the test head.

It would be advantageous to be able to provide balanced compliant motion without requiring all axes of rotation to pass through the center of gravity. In FIG. 9( a ), there are two balanced compliant actuator assemblies 930 and 932 of the same type as shown in FIG. 8( a ) and previously described to eliminate the pitch axis 936 and yaw axis 934 for balancing. The need for smooth movement through the center of gravity.

The two BCAAs 930 and 932 are arranged such that their force bars 112 and links 114 are generally parallel to the roll axis defined by the test head support beam 922. Yaw BCAA 930 is oriented such that a force acting along lever 112 or link 114 will generate a non-zero moment about yaw axis 934. Similarly, pitch BCAA 932 is oriented such that a force acting along force rod 112 or link 114 will produce a non-zero moment about pitch axis 936 . The stationary element 506 of the linear actuator 508 of the two BCAAs 930 and 932 is suitably attached to the test head fixture 926 by means of the bearing 620.

Importantly, the two BCAAs 930 and 932 are able to operate with minimal interaction with each other. Therefore, it is preferable to orient the roll BCAA 930 so that its lever 112 and link 114 act along a line that nearly intersects the pitch axis 936. Similarly, pitch BCAA 932 is also preferably oriented so that its lever 112 and link 114 act along a line that nearly intersects roll axis 934. However, as shown in FIG. 9(a), two force rods 112 and two connecting rods 114 are attached to the test head 100, each with a bearing 116a and 116b, respectively. Therefore, it is impossible to have both the extension axis of the lever 112 and the extension axis of the link 114 of the yaw BCAA 930 intersect the pitch axis 936. In practice, the two rods 112 and 114 are generally close together such that a line parallel to and almost midway between the two rods 112 and 114 intersects the pitch axis 936 . Similarly, it is not possible to intersect the extension axis of the moment bar 112 and the extension axis of the link 114 of the pitch BCAA 932 with the roll axis 934 . In practice, the two rods 112 and 114 are generally close together such that a line parallel to and almost midway between the two rods 112 and 114 intersects the pitch axis 936 . In an alternative configuration of the BCAA as previously described with reference to FIGS. 8( a ) and 8 ( b ), the test head 100 ends of the force bar 112 and link 114 may be connected together with a suitable bracket 800 . In addition, as previously described with reference to FIG. 8( b ), the brackets 116 a and 116 b coupled to the test head 100 may be connected together so that only one bearing 116 b needs to be connected to the test head 100 . In this case, bearing 116 may lie on a line that intersects the axis in question and is parallel to both force bar 112 and link 114 . Figure 9(b) shows a system with such a single bearing 116b. Fig. 9(b) is otherwise the same as Fig. 9(a).

The operation of the system shown in Figures 9(a) and 9(b) can now be described. In general, it can be assumed that the test head 100 is initially in an unbalanced state and in any position where the lockers 118 of the two BCAAs 930 and 932 are disengaged. In a worst-case scenario, the test head 100 rests against physical constraints and mechanical stops built into the system, that is, there are Significant unbalanced forces at extreme locations that are not fully supported by any part of the two BCAAs 930 and 932. The first purpose is to place the test head 100 in a compliant readiness state (ie, an equilibrium state and a compliant neutral position) with respect to the pitch axis 936 and yaw axis 934 . Lock the lock 118 of each of the two BCAAs 930 and 932. Actuator 508 is then used to position test head 100 to a desired first position away from physical constraints and mechanical stops built into the system, where the unbalanced force is captured by the two BCAAs 930 and Either or both lockers 118 and linear actuators 508 of 932 are fully supported. Furthermore, this desired position should be such that the test head 100 can move sufficiently throughout its compliant range so that the compliant readiness occurs without any constraints from physical constraints and mechanical stops that may be built into the system. A disturbance situation is achieved.

Moving test head 100 to a desired position may be accomplished by manual means, which may include commanding the controller to manipulate linear actuator 508 using button commands. Optionally, the controller may be equipped with appropriate algorithms enabling it to utilize the linear actuator 508 in combination with position feedback in an automated positioning sequence.

With the test head 100 now in the desired position, the compliance ready state can now be achieved. In the desired position, the unbalanced force is shared between the two BCAAs 930 and 932. That is to say, the unbalanced force is decomposed into two component forces, one component acts along the yaw BCAA lever 112 , and the other component acts along the pitch BCAA lever 112 . Axis operations can be performed sequentially. First one axis, then another. Since placing one axis in a compliant ready state may disturb the state previously achieved in the other axis, each axis can be repeatedly manipulated again in turn until the balance and position of both axes are satisfied at the same time until. Using known methods, it is possible to implement a control algorithm that automates the sequence of repeated operations for each axis in turn until the complete compliant readiness is achieved. However, in practice, two trials per axis are usually sufficient.

Which axis to choose to operate on first depends on various criteria. In some applications it is always best to address a specific axis first. In some other applications it may be better to choose it randomly. Also, in other applications it may be desirable to select the primary axis to operate on based on a comparison of the forces acting on the individual BCAAs. For example, it is best to work first on the BCAA on which the larger component of the unbalanced force acts. Thus, the outputs of the two force sensors 120 can be compared, and then the axis with the largest unbalanced force component is selected first. Similarly, you can choose the axis with the smallest component of unbalanced force. In each case, if the two components of the unbalanced force are approximately equal, then one or the other axis can be chosen at random.

The axis chosen to be operated on first is now called the primary axis, and the other axis is called the secondary axis. Now follow the appropriate procedure, eg, Method 1 (Fig. 15(b)) or Method 2 (Fig. 15(c)) described above, to bring the first axis into compliance preparation. While doing so, the locker 118 of the BCAA (930 or 932) of the second shaft remains locked. When the first shaft has achieved the compliance ready state, lock the locker 118 of its BCAA (930 or 932); then, the second shaft can also be brought into the compliance ready state. The balance established in the first axis may be disturbed a little in the process of bringing the second axis into compliance readiness. Thus, the process of bringing the first axis into compliance readiness can be repeated. Similarly, the process of bringing the second axis into a compliance ready state can also be repeated. Typically, two iterations usually achieve a reasonable enough degree of compliance readiness in both axes simultaneously.

We can now consider the entire operation of the test head manipulator comprising two BCAAs 930 and 932 as shown in Fig. 9(a). Starting with the test head 100 in any position and not docked with any test equipment, the shaft with the BCAA's 934 and 936 will first enter the compliant ready state while the locker 118 is locked. Then, using the BCAA linear actuator 508 at the appropriate time, the test head 100 is maneuvered into the docking ready position, that is, in the vicinity of the test device with which the test head 100 is to be docked so that the mating parts of the docking device approach the joint. Location. In the docking mechanism shown in Figure 9(a), this would be the point where the guide pin 921 is nearby but not yet inserted into a mating guide hole (not shown). Since the test head 100 is assumed to be in a new orientation, it can further be assumed that it is no longer ready for balance and compliance. Then, the shafts with BCAA (934 and 936) are again put into simultaneous compliance readiness. Now, the BCAA lock 118 is disengaged and the test head 100 is mobilized so that the docking actuation mechanism becomes engaged. In the test head 100 shown in Figure 9(a), this occurs after the guide pin 912 has been inserted into the guide hole (not shown) and the cam follower on the mating mating assembly (not shown) has been fully inserted into the mating cam 910 , when enabling the docking processor 914 to operate. Now activate the docking actuator mechanism and pull the test head into the fully docked position. In FIG. 9( a ), this is accomplished by rotating the docking handler 914 . Other docking actuation mechanisms are known, including electric, pneumatic, and vacuum operated types, and may be used equally well. While docking, the test head 100 is free to make compliant movements if necessary. The test head 100 may remain docked for significant time intervals, and the differential air pressure in the pneumatic cylinder 128 may be lost to leaks. Then, to disengage, keep latch 118 locked and withdraw test head 100 along a straight path perpendicular to the test interface surface (not shown) so that full compliance is not required. Optionally, the system can be equipped with pressure sensors (606a, 606b) so that the pressure on each side of the piston 130 can be recorded by a controller (not shown) when a compliant readiness state has been achieved just prior to docking. Control then uses this information to restore compliance readiness prior to disengagement, and disengagement is no longer associated with compliance motion. Once the test head 100 has been disengaged, the BCAA controlled shafts 934 and 936 can be brought into compliance readiness again as desired.

Figures 10a, 10b, 10c, 10d, 10e, and 10f show a number of embodiments of the invention applied to vertical axes.

Figures 10d and 10e are similar. Both of these have a fixed vertical column 1000 and the actuator drives the vertical axis determined by the column 1000 . In Figure 1Od, BCAAs of the type shown in Figure 8a are used to achieve a state of balance and compliance; however, in Figure 1Oe, BCAAs of the type shown in Figure 6a are used for this purpose. The following discussion relates to what these two diagrams have in common. Where necessary, specific differences based on BCAA type are noted.

Referring generally to Figures 10d and 10e, the main arm 1030 is slidably attached to the post 1000 in the conventional manner utilizing rails 1058 and linear guide bearings (not shown) or round shafts (not shown) and bearings (not shown) . A test head (not shown) is coupled to main arm 1030 by any number of well-known means (also not shown). Thus, the main arm 1030 supports the load. Vertical movement of the main arm 1030 is limited by an upper stop 1032 .

A stationary element 506a comprising a linear actuator 508a of an actuator motor 510a is attached to the base plate 1003 . This actuator 508a is referred to as the main linear actuator 508a. The drive member 514a of the main linear actuator 508a is a screw, typically a ball screw or a trapezoidal screw with suitable pitch and friction so that it cannot be driven backwards. Whether powered or not, the primary linear actuator 508a must be able to safely support the entire load. The primary linear actuator 508a drives the primary arm 1030 through its entire vertical extent so that the travel of the primary linear actuator 508a is equal to the vertical travel of the manipulator. In practice, suitable actuators are commercially available with travels up to 30 inches, and manipulators of this type can accordingly be designed with vertical travels up to 30 inches. For longer vertical travels, telescoping columns of the type shown in Figures 10a, 10b, and 10c can be used.

Bracket 1040 is attached to the top end of main actuator screw 514a. A BCAA of the type shown in FIG. 6a (as in FIG. 10e ) or FIG. 8a (as in FIG. 10d ) couples the main arm 1030 to the bracket 1040 . The BCAA includes a BCAA actuator 508b that precisely positions the lock 118 relative to the force rod 112 . Operation of the main actuator 508a causes vertical movement of the carriage 1040, which in turn causes corresponding vertical movement of the main arm 1030 and its payload. The weight of the main arm 1030 and its payload is transferred to the bracket 1040 and actuator screw 514a through the BCAA mechanism.

Under normal use, the main arm 1030 will first be in a compliant ready state with the lock disengaged. That is, the lock 118 is in a compliant neutral position relative to the force rod 112, and the gas pressure in the pneumatic cylinder 128 is adjusted such that the connecting rod 114 and piston 130 bear substantially the entire load. In this case, the force measured by force sensor 120 is almost zero. Also, the piston 130 is preferably in a compliant neutral position relative to the cylinder 128 so that the load can move through its full compliant range without the piston 130 having to reach the end of its stroke. The main linear actuator 508a is then used to position the load in the vertical direction.

In the case of a vertical axis, the direction of the unbalanced force is always downwards. Thus, there is no need for a double-acting pneumatic cylinder (as shown generally in Figures 6a and 8a); a single-acting pneumatic cylinder 128 with a single gas inlet 132 is sufficient. Thus, as shown, there is only one three position valve 604 , one accumulator 608 , and one optional pressure sensor 606 , which are coupled to the high pressure gas supply 602 via flexible conduit 1010 . The flexible conduit 1010 allows the pneumatic cylinder 128 and its associated parts to move relative to the normally stationary gas supply 602 . The diameters of cylinder 128 and piston 130 are preferably sufficient to support main arm 1030 and the load attached thereto. For example, if the combined weight of the main arm 1030 and attached load is 1000 pounds, and the maximum gas pressure that can enter the cylinder 128 is 100 pounds per square inch, then the cylinder 128 and piston 130 preferably have an area of at least 10 square inches, and The corresponding diameter is at least approximately 3.57 inches. The required cylinder 128 diameter is proportional to the square root of the load. Thus, doubling the load requires increasing the diameter of the cylinder 128 by the square root of two, or, alternatively, doubling the gas pressure (eg, with the aid of a gas pressure doubler). If the required cylinder 130 size is too large for a given load, then two or more smaller cylinders 128 may be arranged in parallel.

Generally, it is best to have a vertically compliant motion of about ±1 inch. In general, it is desirable to have the stroke of piston 130 slightly greater than the compliance range determined by compliance stops 522 and 524 so that the piston does not "bottom out" with locker 118 against the compliance stops. The system is designed such that, during operation, if the lock 130 is in its compliant neutral position relative to the extended compliant stop 522 and the retracted compliant stop 524, then the piston 130 is in the neutral position relative to the cylinder 128. Soft neutral position. During docking, the load can move over the entire compliance range as defined by compliance stops 522 and 524 .

In Figure 10(d), cylinder 128 and locker 118 are rigidly attached to each other by bracket 1052, and the system can be conveniently arranged so that whenever locker 118 is in the compliant neutral position, piston 130 is also in the neutral position. In such a position. For example, the preferred compliant neutral position tends to be the centered position; and the system can be arranged so that when the lock 118 is centered between the compliant stops 522 and 524, the piston 130 is at the midpoint of its stroke. A single position sensor 1046 shown in Figure 10(d) is sufficient to position both the lock 118 and the piston 130 in their respective compliant neutral positions.

In FIG. 10e , the lock 118 is not attached to the cylinder 128 , but the lock 118 can and does move relative to the cylinder 128 . To achieve the desired goal of positioning both the locker 118 and the piston 130 in a compliant neutral position, information in addition to that provided by the illustrated position sensor 1046 is required. Positional information about the piston 130 or BCAA actuator 508b is sufficient. In the preferred embodiment, the system is designed so that if the BCAA actuator 508b is fully extended and the lock 118 is in its compliant neutral position, the piston will also be in its compliant neutral position. Detecting the fully extended state of the BCAA actuator 508b is readily accomplished in known ways, including using limit switches (not shown) included in purchased BCAA actuator 508b assemblies, or by operating The BCAA actuator 508b is sufficient for it to reach the end of its swim and the time interval between stops is achieved. In the latter case, it is recommended to use a current limited voltage supply for the BCAA actuator motor 510b to minimize the possibility of overheating. An alternative method is to detect the position of the piston 130 relative to its cylinder 128 . Pneumatic cylinder-piston assemblies are often equipped with limit switches (not shown in the figure) by their manufacturer, which will allow this detection.

In some applications, it may be desirable to have the BCAA actuator 508b able to position the lock 118 over the entire compliant range defined by the compliant stops 522 and 524 when the piston is at any point in its range of travel. In this case, the range of the BCAA actuator 508b in FIG. 10e is at least the sum of the compliance range plus the stroke of the piston 130 . Furthermore, in Fig. 10e, it is necessary to directly detect the position of the piston 130 during the process of achieving the compliance ready state. In contrast, the required range of the BCAA actuator 508b in Figure 10d is at least the greater of the compliance range or the stroke of the piston 130, and only a single position sensor 1046 as shown is sufficient.

The sequence of operations to achieve a compliant ready state is as follows:

First, we define a "balanced sequence".

Balance sequence (as shown in Figure 15(d))

1. As shown in step 1530, lock the locker 118.

2. As shown in step 1531, adjust the gas pressure in the cylinder 128 until the force sensor 120 indicates that the force exerted by the force rod 112 on the lock 118 is less than a threshold.

Now, in order to achieve the compliant ready state ("compliant ready sequence 1")

Compliant preparation sequence 1 (shown in Figure 15(e)) (applicable to both Figures 10d and 10e)

1. As shown in step 1540, a balancing sequence is performed.

2. As shown in step 1541, the lock 118 is unlocked.

3. As shown in step 1542a, use the BCAA actuator 508b to position the lock 118 in the compliant neutral position.

4. If the system is of the type shown in Figure 10e, lock the locker 118 as shown in step 1542b, then, as shown in an alternative embodiment in step 1542c, use the BCAA actuator 508b to position the piston 130 in the Its supple neutral position.

5. As shown in step 1543, if the force sensor 120 indicates that the balance has been disturbed or deteriorated (eg, if the measured force is greater than the force threshold), then the balancing sequence is repeated.

6. As shown in step 1544, if the lock 118 is locked, then unlock the lock 118.

It should be noted that in compliance preparation sequence 1, preferably the force threshold for balancing must be less than the sum of the static friction force and the pneumatic piston breakaway force. Otherwise, when the lock 118 is released in step 1541, the lock 118 will move against the compliant stop 522 or 524, and operation of the BCAA actuator 508b will not cause the lock 118 to act against the force Movement of rod 112 . However, the combined force of static friction and air piston breakaway force is usually in the range of 5 to 10 pounds or a little more, so it can be balanced within 1 to 3 pounds.

For reasons as described later, preferably the system is designed to satisfy the aforementioned criteria. However, in cases where achieving this criterion is impossible or impractical, there are other possible alternative sequences to achieve a compliant readiness state. Two sequences are now described.

A first alternative sequence uses the pneumatic cylinder 128 and piston 130 to position the piston 118 by means of the following sequence of operations ("Compliance Preparation Sequence 2"):

Compliant preparation sequence 2 (shown in Figure 15(f)) (applicable to both Figures 10d and 10e)

1. As shown in step 1550, the lock 118 is unlocked.

2. As shown in step 1551 , exhaust gas from the cylinder 130 so that the extended compliant stop 522 rests on the lock 118 .

3. As shown in step 1552, adjust the BCAA actuator 508b so that it is near the midpoint of its stroke or near its desired compliant neutral position (this step requires the use of well known solutions including the use of the position sensor 1046 to detect the actuation The position of the driving part 514b of the actuator 508b relative to its stationary part 506b. In some cases, the actuator 508b is equipped with a limit switch that detects when the driving part 514b reaches the end of its travel; to achieve the desired result).

4. As shown in step 1553, inject gas into the cylinder 128, increasing the cylinder gas pressure to the point where the piston 130 breaks free and moves upward.

5. As shown in step 1554, while monitoring the position of the force rod 112 relative to the lock 118 by means of the position sensor 1046 as shown, continue to adjust the gas pressure of the cylinder 128 so that the piston 130, and thus the connecting rod 114 and force Rod 112 continues to move.

6. As shown in step 1555a, when the force rod 112 reaches the compliant neutral position relative to the locker 118, lock the locker 118 and stop adjusting the gas pressure in the cylinder 128.

7. As shown in an alternative embodiment in step 1555b, if the system is of the type shown in Figure 10e, then use the BCAA actuator 508b to position the piston 130 in its compliant neutral position.

8. As shown in step 1556, a balancing sequence is performed.

The second alternative sequence is to change the design slightly. Contrary to the preference to avoid bottoming of the piston 130 as described above, the system is arranged so that the piston 130 rests against the lower end of the cylinder 128 if there is no gas pressure in the cylinder 128 and the lock 118 is disengaged. In fact, the piston 130 resting against the bottom end of the cylinder 128 performs the function of extending the compliant stop 522, and thus, the extending compliant stop 522 can be omitted from the system (note also that if the piston 130 is placed against the At the top end of the cylinder 128 as a stop, the retracted compliant stop 524 can then be omitted). Then, the following "Compliant Prep Sequence 3" can be applied.

Compliant preparation sequence 3 (shown in Figure 15(g)) (applicable to both Figures 10d and 10e)

1. As shown in step 1560, retract actuator 508a to a position where the piston in cylinder 128 has reached the bottom of its stroke (depending on the design, this may be controlled by a limit switch incorporated in actuator 508a, via Actuator 508a is driven in the retracted direction long enough to achieve the fully retracted state, or other known techniques).

2. As shown in step 1561, the lock 118 is unlocked.

3. As shown in step 1562, the gas is exhausted from the cylinder 128 so that the piston 130 rests on the bottom of the cylinder 128.

4. As shown in step 1563, using the BCAA actuator 508b, position the lock 118 in the compliant neutral position.

5. As shown in step 1564, a balancing sequence is performed.

6. Keeping the lock 118 locked, use the BCAA actuator 508b to lift the load to a position where the piston 130 is in a compliant neutral position relative to the cylinder 128 (typically near midpoint), as shown in step 1565.

7. As shown in step 1566, due to the slight imbalance caused by the movement of the piston 130, the balancing sequence is performed again.

In compliance preparation sequences 2 and 3, it is not guaranteed that the equilibrium sequence threshold is not less than the static friction combined with the breakaway force. Therefore, the lock 118 should remain in the locked state until the ready-to-dock position is achieved and the load is partially supported by the docking device, (or other desired state). However, it should be appreciated that if the lock 118 can be left disengaged while the load is positioned vertically by means of the primary actuator 508a, then the position sensor 1046 can be used to detect whether the load encounters any obstructions. If an obstacle is encountered, then the associated force on the load will move the force bar 112 relative to the lock 118 . This movement may be detected by the position sensor 1046, thereby providing a signal to take appropriate action, such as stopping the main actuator 508a. The system is then preferably designed such that the compliant preparation sequence 1 can be used reliably.

FIG. 10f can also be applied to fixing vertical columns 1000 . The vertical axis is driven by actuator 508 . BCAAs of the type shown in Figure 8(a) and Figure 10d provide vertical actuation, as well as balanced and compliant states. Many of the things described in the description of Figures 10D and 10E are also applicable to Figure 10F, so in this section we mainly focus on what is new and what is different.

The system in Figure 1Of contains a single linear actuator 508, which is equivalent to the linear actuator 508 of the BCAA in Figure 8a. Also, it is equivalent in travel and specification to the main linear actuator 508a in Figures 1Od and 1Oe. The distal end of the actuator screw 514 is attached to the lock 118 . Bracket 1094 attaches lock 118 to single-acting pneumatic cylinder 128 . The main arm 1030 is supported by the subassembly of the connecting rod 114 and piston 130, the subassembly of the force rod 112, the force sensor 120, and the lock 118, or a combination of these two subassemblies depending on the system state. The device attached to the distal end of the linear actuator 508 and coupled to the main arm 1030 is equivalent to the balancing mechanism of FIG. 1( a ).

In order to bring the main arm 1030 into a compliant ready state, the following sequence ("compliant ready sequence 4") may be used:

Compliant preparation sequence 4 (as shown in Figure 15(h))

1. As shown in step 1570, the lock 118 is unlocked.

2. Exhaust gas from cylinder 128 so that extended compliant stop 522 rests against lock 118 as shown in step 1571 .

3. As shown in step 1572, inject gas into the cylinder 128, increasing the cylinder gas pressure to the point where the piston 130 breaks free and moves upward.

4. As shown in step 1573, while monitoring the position sensor 1046, continue to adjust the cylinder gas pressure to continue to move the piston 130, and thus the connecting rod 114 and force rod 112.

5. As shown in step 1574, when the force rod 112 reaches the compliant neutral position relative to the lock 118, lock the lock 118 and stop adjusting the gas pressure in the cylinder 128.

6. As shown in step 1575, a balancing sequence is performed.

If the equilibrium sequence force threshold is less than the resultant of the static friction force and the breakaway force of the piston 130, then the lock 118 can be disengaged since the actuator 508 is used to drive the main arm 1030 and the load it supports to the desired vertical position, And keep it unwrapped. As discussed earlier, this is preferred. Otherwise, the lock 118 should be maintained in the locked condition until the ready docked position (or other suitable desired condition) in which the load is partially supported by the docking device is achieved.

Both Figures 10a and 10b have a manipulator, and a telescoping column 1000 similar to that described in PCT Application PCT/US01/06456 to Holt et al. In Figure 10a, BCAAs of the type shown in Figure 8a are used to provide a state of balance and compliance; while in Figure 10b, BCAAs of the type shown in Figure 6a are used for this purpose. The following discussion relates to what these two diagrams have in common. Where necessary, specific differences based on BCAA type are indicated.

The main arm 1030 is slidably attached to the upper section 1001 of the telescoping column 1000 using linear guides 1058 and linear guide bearings (not shown). This provides about ± 1 inch of vertical inching of the main arm 1030 relative to the upper section 1001 . Upper and lower main wall stops 1032 and 1054 respectively limit the movement of main arm 1030 relative to upper section 1001 . A test head (not shown) is coupled to main arm 1030 by any number of well-known means (also not shown). Thus, the main arm 1030 supports the load.

The first linear actuator 508a moves the middle section 1002 of the column in a vertical direction relative to the fixed lower section 1004 and the base 1003 . The second linear actuator 508b is mounted on the first bracket 1034 attached to the middle section 1002 and it moves the upper section 1001 in a vertical direction relative to the fixed middle section 1002 . Telescoping of the telescoping column 100 provides a first vertical positioning of the load. During the docking process, the slight movement of the main arm 1030 between the upper stop 1032 and the lower stop 1054 is used for compliant vertical movement.

The BCAA 508c is coupled between the second bracket 1040 attached to the upper section 1001 and the main arm 1030. A BCAA 508c of the type used in Figure 8a is used in Figure 10a, and a BCAA 508c of the type used in Figure 6a is used in Figure 10b. In each case, the BCAA 508c is used to control vertical inching and provide compliance over the entire range of vertical inching.

The compliant readiness in this system is similar to that of a system containing fixed posts 1000 as in Figures 1Od and 1Oe. In addition, however, the main arm 1030 also needs to be positioned at a specific approximately concentric position between the upper main arm stop 1032 and the lower main arm stop 1054 (or, in the absence of the lower stop 1054, the positioning to a position at a specified distance relative to the upper stop 1032). We call that position of the main arm 1030 the compliant neutral position.

Then, the state about the compliant readiness can be stated as follows:

1. The gas pressure in the pneumatic cylinder 128 must be such that the connecting rod 114 and piston 130 bear substantially the entire load of the main arm 1030 and its load.

2. The lock 118 must be in a compliant neutral position (typically, midway between the two) with respect to the extended compliant stop 522 and the retracted compliant stop 524,

3. The piston 130 is in a compliant neutral position relative to its cylinder 128, and

4. The main arm 1030 must be in a compliant neutral position relative to the upper main arm stop 1032 and the lower main arm stop 1054 (typically, midway between the two).

Achieving co-location of the locker 118 and piston 130 in items 2 and 3 of the upper list in Figures 10a and 10b requires the same as previously described with reference to Figures 10d and 10e. In particular, since the locker 118 is attached to the cylinder 128 by the bracket 1052, this can be achieved with the design in Figure 10a, and requires positional information of the BCAA actuator 508c or piston 120 in Figure 10b.

To provide the general solution needed to co-locate the main arm 1030 with the combination of the lock 118 and the piston 130, a position sensor 1056 detecting the position of the main arm 1030 relative to the upper section 1001 is shown in Figures 10a and 10b. It should be noted that in FIG. 10b (not in FIG. 10a ), this sensor 1056 is also effectively indicative of the position of the piston 130 relative to the cylinder 128 . Thus, in both Figures 10a and 10b, sensor 1056 combined with position sensor 1046 as shown is sufficient to provide all necessary position sensing for a general solution.

Alternatively, and given overall cost considerations, the system is preferably designed so that when the BCAA actuator 508c is fully extended, the main arm 1030 is in its compliant neutral position and the lock 118 is also in its compliant position. neutral position. For this constraint, the master arm position sensor 1056 can be omitted from Figures 10a and 10b. However, it is necessary to include means to detect when the BCAA actuator 508c is fully extended. A typical solution to this has been discussed with reference to Figures 10d and 10e.

The following sequence of operations can be used to position the device in the compliance ready state ("Compliant Ready Sequence 5")

Compliant preparation sequence 5 (as shown in Figure 15(i))

1. As shown in step 1580, a balancing sequence is performed.

2. As shown in step 1581a, unlock the lock 118 and position the lock 118 in the compliant neutral position using the BCAA actuator 508c.

3. As shown in the alternative embodiment in step 1581b, if the system is of the type shown in FIG. sexual position.

4. As shown in step 1582, if the main arm 1030 is not already in its compliant neutral position (e.g., by design), then lock the lock 118 and move the main arm 1030 using the BCAA actuator 508c to its supple neutral position.

5. As shown in step 1583, if the force sensor 120 indicates that the force in the force rod 112 is greater than the force threshold, then a balancing sequence is performed.

6. As shown in step 1584, if the lock 118 is locked, then it can be unlocked.

It should be noted that steps 1581 and 1584 require the force threshold to be less than the sum of the static friction force and the breakaway force of the piston 130 . As in the case of Figures 1Od and 1Oe, this is the preferred criterion for the system. In this regard, as the main arm 1030 and the load it supports are driven to the desired vertical position, the latch 118 can be released and remain in the released state. Additionally, as described in PCT Application PCT/US01/06456 to Holt et al., one of the two position sensors 1046 and 1056 may be used to detect movement of the main arm 1030 relative to the upper section 1001, which is indicative of an obstacle being encountered.

In the case where the system is set so that the force threshold is not necessarily smaller than the resultant force of the static friction force and the breakaway force of the piston 130, the first three steps of the compliance preparation sequence 5 can be replaced by compliance preparation sequence 2 or compliance preparation sequence 3 . The resulting sequences are provided below. It should be noted that in this case, it is generally necessary to keep the lock 118 locked while positioning the load, and that the position sensors 1046 and 1056 cannot be used to detect moving obstacles. However, in any case, the force sensor 120 can be used to detect moving obstacles.

Compliant Prep Sequence 6 (shown in Figure 15(j)) (applicable to both Figures 10a and 10b, where pneumatic cylinder 128 is used in positioning locker 118 and balance threshold is greater than static friction and breakaway forces)

1. As shown in step 1590, the lock 118 is unlocked.

2. Exhaust gas from cylinder 128 so that extended compliant stop 522 rests against lock 118 as shown in step 1591 .

3. As shown in step 1592, adjust the BCAA actuator 508b so that it is near the midpoint of its stroke or near its desired compliant neutral position (this step requires the use of well known solutions including the use of the position sensor 1046 to detect the actuation The position of the driving part 514b of the actuator 508b relative to its stationary part 506b. In some cases, the actuator 508b is equipped with a limit switch that detects when the driving part 514b reaches the end of its travel; to be used in combination to obtain the desired result).

4. As shown in step 1593, inject gas into the cylinder 128, increasing the cylinder gas pressure to the point where the piston 130 breaks free and moves upward.

5. As shown in step 1594, while monitoring the position of force rod 112 relative to lock 118 by means of position sensor 1046 as shown, continue to adjust the gas pressure of cylinder 128 to make piston 130, and thus connecting rod 114 and The force rod 112 continues to move.

6. When the lever 112 reaches the compliant neutral position relative to the lock 118, lock the lock 118 and stop adjusting the gas pressure in the cylinder 128, as shown in step 1595a.

7. As shown in an alternative embodiment in step 1595b, if the system is of the type shown in Figure 10b, then the piston 130 is positioned in its compliant neutral position using the BCAA actuator 508b.

8. As shown in step 1596, if the main arm 1030 is not already in its compliant neutral position (e.g., by design), then lock the lock 118 and move the main arm 1030 using the BCAA actuator 508c to its supple neutral position.

9. As shown in step 1596, a balancing sequence is performed.

Compliant Preparatory Sequence 7 (shown in Figure 15(k)) (applicable to both Figures 10a and 10b, where the piston 130 can rest on the bottom of the cylinder 128 and the equilibrium threshold is greater than the static friction and breakaway forces)

1. As shown in step 1600, retract actuator 508c to a position where piston 130 in cylinder 128 reaches the bottom of its stroke (depending on the design, this may be controlled by a limit switch incorporated in actuator 508a, via Actuator 508a is driven in the retracted direction long enough to achieve a fully retracted state, or by other known techniques).

2. As shown in step 1601, unlock the locker 118 .

3. As shown in step 1602, the gas is exhausted from the cylinder 128 so that the piston 130 rests on the bottom of the cylinder 128.

4. As shown in step 1603, using the BCAA actuator 508b, position the lock 118 in the compliant neutral position.

5. As shown in step 1604, lock locker 118 and lift load using BCAA actuator 508c to a position where piston 130 is in a compliant neutral position relative to cylinder 128 (typically near midpoint).

6. As shown in step 1605, if the main arm 1030 is not already in its compliant neutral position (e.g., by design), then use the BCAA actuator 508c to move the main arm 1030 to its compliant neutral position superior.

7. As shown in step 1606, a balancing sequence is performed.

Figure 10c is also a manipulator comprising a telescoping column 1000 similar to that described in PCT Application PCT/US01/06456 to Holt et al. As previously described, main arm 1030 is slidably attached to upper section 1001 using linear guides 1058 and linear guide bearings (not shown). This provides about ± 1 inch of vertical inching of the main arm 1030 relative to the upper section 1001 .

The first linear actuator 508a moves the middle section 1002 of the column 1000 in a vertical direction relative to the fixed lower section 1004 and the base 1002 . A BCAA of the type shown in Figure 8a is suitable for moving the upper section 1001 relative to the middle section 1002 in the direction perpendicular to the axis of fretting and providing a balanced and compliant state.

The stationary element 506b of the second linear actuator 1042 is attached to the first bracket 1040 which is in turn attached to the midsection 1002 . The distal end of the actuator 514b is attached to the BCAA lock 118 . The lock 118 is attached to a second bracket 1076 which is also attached to the upper section 1001 . Thus, the linear actuator 508b drives the vertical movement of the upper section 1001 relative to the middle section 1002 . The pneumatic cylinder 128 of the BCAA is also attached to the second bracket 1076 . The force rod 112 and the connecting rod 114 are used to support the main arm 1030 . It can be seen that the equipment comprising the second linear actuator 508b, the pneumatic cylinder 128, the piston 130, the connecting rod 114, the force rod 112, the force sensor 120, the position sensor 1056, the compliant stoppers 522 and 524, etc. That type of BCAA is shown, with the added feature that bracket 1076 is also attached to movable upper section 1001.

The conditions for the master arm 1030 to be in a compliance ready state are the same as those of Figures 10a and 10b. However, the situation is simpler since the lock 118 cannot be positioned independently of the upper section 1001 as in Figures 10a and 10b. The end result is that if the main arm 1030 is in its compliant neutral position, then the lock must also be in its compliant neutral position, and vice versa. The following sequence can be used to achieve a compliance ready state ("Compliant Ready Sequence 8")

Compliant preparation sequence 8 (as shown in Figure 15(l))

1. As shown in step 1610, the lock 118 is unlocked.

2. As shown in step 1611, exhaust gas from the cylinder 128 so that the main arm 1030 rests on the lower main arm stop 1054 and/or the extended compliant stop 522 rests on the second bracket 1076 (note , in which case both the second bracket 1076 and the lower main arm stop 1054 are secured to the upper section 1001).

3. As shown in step 1612, inject gas into the cylinder 128 to increase the cylinder gas pressure to the point where the piston 130 breaks free and moves upward.

4. As shown in step 1613, while monitoring the master arm position sensor 1056, continue to adjust the cylinder gas pressure to continue to move the piston 130, and thus the connecting rods and force rods 114 and 112, respectively, and the master arm 1030 upwardly.

5. As shown in step 1614, when force rod 112 and main arm 1030 reach their compliant neutral position, lock lock 118 and stop adjusting the gas pressure in cylinder 128.

6. As shown in step 1615, a balancing sequence is performed.

If the equilibrium sequence force threshold is less than the resultant of the static friction force and the breakaway force of the piston 130, then the lock 118 can be disengaged since the actuator 508 is used to drive the main arm 1030 and the load it supports to the desired vertical position, And keep it unwrapped. Otherwise, the lock 118 should be maintained in the locked condition until the ready-to-dock condition (or other desired condition) in which the load is partially supported by the docking device is achieved.

As can be seen from the above, one position sensor 1056 is sufficient in this case.

FIG. 11 shows a few of the various exemplary embodiments of the present invention used in a single test head manipulator 1101 . Manipulator 1101 in FIG. 11 contains base 1103 as described in U.S. Patent Application 09/646,072 issued to Holt et al., fixed post 1000, and supports a test head 100 of the type previously described in FIG. 9( b). main arm 1030 of the main arm 1030 (note that the test head 100 in Figure 11 is shown in cross-section, and for simplicity, the docking hardware shown in Figure 9(b) is not repeated in Figure 11 - as well as the internal pitch and roll rotating mechanism). The manipulator 1101 has seven axes of motion, three axes of translation and four axes of rotation. The translational axes are medial-outer axis 1128 , left-right axis 1126 , and up-down (or vertical) axis 1130 . The rotation axes are the yaw axis 1115 which rotates around the base, the pitch axis 1113 , the roll axis 1117 , and the roll axis 1111 which rotates around the test head 100 . Typical motion requirements are ±10 inches inside and outside, ±5 inches side to side, ±30 inches vertical, 30+ degrees yaw, ±95 degrees roll, ±5 degrees pitch, and ±5 degrees yaw. In this embodiment, movement of all manipulator axes is powered by motion drive actuators. The actuators are controlled by a central controller (not shown) to move the test head 100 from one position and orientation to another within the manipulator 1101 motion envelope. Any of a number of techniques that implement control over several axes to control motion can be used. Use appropriate position sensors on all axes to provide feedback to the controller.

In this manipulator 1101 , the cable pivot ring assembly 924 is mounted on the end of a bracket 1114 attached to the main arm 1030 . Test head 100 is supported by test head support beam 922 secured to test head mount 926 . Test head mount 926 is coupled to rotatable ring 1102 of cable pivot ring 924 such that the centerline of support beam 922 passes through the center of rotatable ring 1102 and is perpendicular to the plane of rotatable ring 1102 . The apparatus is configured to enable the test head 100 to rotate ±95 degrees about the axis defined by the centerline of the support beam 922 , ie, the roll axis of motion 111 .

In FIG. 11 , the roll axis 1111 passes through the center of gravity of the test head 100 so that the roll is substantially balanced and free of effective weight. Accordingly, movement of the test head 100 about the roll axis can be powered by means of a motor motor-gearbox-clutch arrangement as described in US Patent Application Serial No. 09/646,072 to Holt et al. When the motor (not shown) is not being driven, the clutch is disengaged, allowing balanced and compliant motion about the roll axis 1111.

The pitch axis 1113 and the yaw axis 1117 of the test head 100 do not pass through the center of gravity of the test head 100 . A pitch BCAA 932 and a roll BCAA 930 as shown in FIG. 9( a ) and previously discussed are coupled between the test head 100 and the test head mount 926. These are shown schematically in Figure 11 as mechanism 1106 and mechanism 1112, respectively. The details are the same as in Fig. 9(a). The two BCAAs 1106 and 1112 are used to position the test head 100 relative to its pitch and yaw axes, respectively, and to provide balanced, compliant motion for docking or/or manual manipulation. A range of movement of ±5 degrees is provided in the two pitch and yaw axes 1113 and 1117, respectively.

Vertical movement of the test head 100 may be provided by an arrangement including a BCAA as previously discussed for the alternative arrangements shown in Figures 10(d), 10(e) and 10(f). In Fig. 11, the arrangement of Fig. 10(e) is shown explicitly as mechanism 1118 (note that the flexible gas conduit and high pressure gas supply are not shown in Fig. 11). It can be seen that if a telescoping column is used in place of the fixed column 1000, then one of the types of arrangements shown in Figures 10(a), 10(b), or 10(c) could be used. The system in FIG. 11 is designed such that the static friction force plus the breakaway force of the pneumatic cylinder 128 is greater than the balance force threshold so that the test head 100 can be positioned vertically with the lock 118 unlocked.

The deflection motion in the base 1103 is in the horizontal plane and is achieved by a compliant drive mechanism shown schematically as mechanism 1120 in FIG. 11 . Generally, motion along this axis is substantially balanced and free from external forces. In this case, a compliant drive mechanism of the type shown in any of Figures 5(a), 5(b), 5(c) or 7 could in principle be used. Figures 5(a) and 7 are preferred since a deflection movement of about 30+ degrees requires a sufficiently long stroke actuator. However, in some applications, a thick, slightly elastic test head cord (not shown) connecting the test head 100 to the test cabinet (not shown) may apply a force along this axis that affects the motion. Force that varies with position. In this case, the BCAA can be used as mechanism 1120, providing actuation, and balancing the compliant state. A BCAA configuration as shown in Figure 6(b) can be used. Furthermore, BCAAs of the type shown in Figure 8(a) or 8(b) may be similarly suitable for this application.

Similarly, the in-out and left-right motions of the base 1103 are also in the horizontal plane. The compliant drive mechanisms associated with these axes are also schematically represented in Figure 11 as mechanism 1122 and mechanism 1124, respectively. As with deflection motion, motion along these axes is essentially balanced and bears no external force. The required range of motion in these axes is typically from ±5 inches to ±10 inches. Therefore, it is best to choose the mechanism of Figure 5(a) or Figure 7 to provide the required stroke. BCAA mechanisms of the type shown in Figures 6(a), 6(b), 8(a), or 8(b) may be employed and utilized where significant variable rope forces exist in the system, which Good for the horizontal axis.

It should be noted that the BCAA and compliant drive mechanism shown in FIG. 11 are shown in view for clarity of description and discussion. In practical devices, certain mechanisms or parts of them are invisible from a given viewing angle. For example, components of mechanisms 1120, 1122, and 1124 are typically located among various plates, including manipulator base 1103, along with other components. Also, for example, mechanism 1118 may be on one side or the other side or back of manipulator post 1105 .

Information about docking

An exemplary embodiment of an actuator-driven docking is explained with reference to FIG. 11 . In FIG. 11 , manipulator 1101 is used to prepare the test head for a docked position, and then a separate, independent docking actuator (not shown) draws test head 100 into the final or fully docked position. Various types of guide mechanisms, eg, guide pins and holes, kinematic mechanisms, etc., are known to ensure proper and precise alignment of test head 100 with a device handler/detector (not shown). The test head 100 must move compliantly relative to the manipulator 1101 when operating the actuator. A mechanism, which may be an actuator mechanism, securely latches the test head in a certain position when docking is complete.

As test heads 100 have grown larger, the basic idea of the '815 dock has evolved to a dock containing three to four sets of guide pins 912 interconnected by cords 915, guide pin receptacles 912a, and circular cams 910. Figures 12(a), 12(b), 12(c), and 12(d) of the present application show a combination of four guide pins 912 and small holes (receptacles) 912a and four circular cams 910 for docking pieces. Although such a four-point dock is configured to include an actuator handle attached to each of the four cams 910 , the illustrated dock also includes a single actuator handle 914 attached to a cord drive 917 . This arrangement places the single actuator handle in a convenient position for the operator when the cord driver 917 is rotated by the docking handle 914 . Furthermore, by properly adjusting the ratio of the diameter of the cam 910 to the diameter of the cable drive, greater mechanical advantage can be obtained.

Figures 12(a) to 12(d) are now considered in more detail. Figure 12(a) shows in perspective the test head 100 fixed in the frame 200, which in turn is supported by a manipulator (not shown). Also shown is a cut-away section of device handler 1208 with which test head 100 may interface. Figure 12(b) shows device processor 1208 on a slightly larger scale and in more detail (again, the reader is reminded that the term "processor" is used here without loss of generality to include plug-in device processors, on-chip Any of various test equipment including detectors, etc.). Looking briefly at the cross-sectional view in FIG. 12( c ), it can be seen that the test head 100 includes an electrical interface 1226 and the device handler 1208 includes a corresponding electrical interface 1228 . Electrical interfaces 1226 and 1228 typically contain hundreds or thousands of tiny, delicate electrical contacts that must be precision-engaged in a manner that provides a reliable corresponding individual electrical connection when the test heads are finally docked. As shown in this exemplary case, the lower surface of the device handler 1208 contains the handler electrical interface 1228, and the test head 100 is generally docked with a bottom-up motion. Other orientations are also possible and known, including: interface with the upper surface by means of downward motion, interface with the vertical surface with the aid of horizontal motion, and interface with a plane at an angle to both the horizontal and vertical planes .

Returning to FIGS. 12( a ) and 12 ( b ), these figures show a complete four-point docking device; Attached to test head 100 is faceplate 1206 . Four guide pins 912 are attached to and near the four corners of panel 1206 . Faceplate 1206 has a central hole and is attached to test head 100 so that test head electrical interface 1226 protrudes through the aperture, and guide pins 912 define an approximate rectangle that is nearly concentric with electrical interface 1226 .

A gusset 1214 is attached to the lower surface of the device handler 1208 . Gusset 1214 has a central hole and is attached to device processor 1208 so that processor electrical interface 1228 protrudes through the hole. Four corner pieces 1216 are attached to the corner plate 1214, one each adjacent each of its four corners. Each gusset 1216 has a guide pin aperture or receptacle 912a drilled therein. Each guide pin aperture 912a is for a respective guide pin 912 . These are arranged so that when mated, each guide pin 912 will be fully inserted into its respective guide pin aperture 912a. Thus, guide pins 912 and guide pin apertures 912a ensure alignment between the test head and the device handler.

Four docking cams 910 are rotatably attached to the panel 1206 . Cam 910 is circular and similar to those described in the '815 patent. In particular, each cam 910 presents a side helical groove around its circumference, and an upper cutout on the upper surface. Each docking cam 910 is positioned adjacent to a respective guide pin 912 such that it is generally centered on a line extending approximately from the center of the test head electrical interface 1226 through the respective guide pin 912 such that the guide pin 912 is positioned on the cam. 910 and the test head electrical interface 1226. The corners of the corner piece 1216 and the corner plate 1214 have circular cutouts so that when the guide pin 912 is fully inserted into the guide pin aperture 912a in the corner piece, the circumference of each cam 910 matches the circular shape in its respective corner piece 1216. The cutouts are contiguous and concentric with the circular cutouts in their respective corner pieces 1216 . This arrangement ensures an initial guided alignment between the docking components when the test head 100 is first manipulated into a docking position for the device handler 1208 .

A circular cord driver 917 with attached docking handle 914 is also rotatably attached to panel 1206 . Docking cables 917 are attached to each of the cams 910 and to a cable driver 917. The pulley 1224 guides the path of the rope to and from the rope drive 917 as appropriate. By applying force to the handle 914, the cord driver 917 can be rotated. As the cord driver 917 rotates, it transfers force to the cord 915 which in turn causes the cam 910 to rotate synchronously.

Extending from the circular cutout of each gusset 1216 is a cam follower 1210 . Cam followers 1210 fit into upper notches on the upper surface of their respective cams 910 . FIG. 12( c ) shows a stage in the process of interfacing the test head 100 with the handler 910 in cross-sectional view. Here, guide pin 912 is partially inserted into guide pin aperture 912a in gusset 1216 . Note that in this exemplary case, the guide pins 912 taper near their distal ends and have a constant diameter closer to their point of attachment to the panel 1206 . In Figure 12(c), the guide pin 912 has been inserted into the guide pin aperture 912a to the point where the region of constant diameter just enters the guide pin aperture 912a. And in FIG. 12( c ), each cam follower 1210 has been inserted into the upper cutout on the upper surface of its corresponding cam 910 until it is at the depth of the uppermost end of the cam helical groove. In this configuration, the abutment is ready to be actuated by applying force to the handle and rotating the cam. Thus, this configuration may be referred to as a "ready to drive" position.

Figure 12(d) shows the result of full rotation of the cam 910 in a cross-sectional view. Test head 100 is now fully interfaced with processor 1208 . It can be seen that the cam 910 has rotated and has brought the cam follower 1210 along the helical groove to a point closer to the panel 1206 . In addition, the guide pins 912 are fully inserted into their corresponding guide pin apertures 912a. It can be seen that the tightness of fit between the constant diameter region of the guide pin 912 and the corresponding guide pin aperture 912a determines the final alignment between the processor electrical interface 1228 and the test head electrical interface 1226 .

In light of the foregoing discussion, it is now appropriate to discuss the docking process more fully and to define certain terms. The purpose of the docking is to precisely mate the test head electrical interface 1226 with the processor electrical interface 1228 . Each electrical interface 1226 and 1228 defines a plane that is typically, but not necessarily, nominally parallel to the top end of the electrical interface. When docked, these two planes must be parallel to each other. To prevent damage to the electrical contacts, it is preferable to first align the two interfaces 1226 and 1228 in five degrees of freedom before bringing the electrical and mechanical contacts into mechanical contact with each other. If, in the mated position, the plane defined by the interface is parallel to the X-Y plane of Figure 14, then alignment must be in the X, Y and ΘZ directions in order for the contacts to align with each other. In addition, the two planes are made parallel by rotation in the θX and θY directions. The process of making two electrical interfaces parallel to each other is called planarization of the "interface"; and, when planarization has been achieved, the interface can be said to be "planarized" or "coplanar." Once planarized and aligned in the X, Y, and θZ directions, alignment continues by causing motion in the Z direction perpendicular to the plane of the processor electrical interface 1228 . During the docking process, the test head 100 is first moved near the handler 1208 . Further movement aligns the circular cutout of the gusset 1216 with the cam 910 first. This position, or the position just before this position, may be considered the "ready to dock" position. More specifically, a "ready-to-dock" position refers to a position in which a first guide alignment device has just been engaged or is about to be engaged. At this stage, depending on design details, the distal ends of the guide pins are prepared to enter their respective guide holes. Further movement will bring the test head into the "ready to drive position" discussed above with respect to Figures 12(a) to (d). More specifically, "ready to drive position" means that the test head has reached a position where it can drive the docking device. In the ready-to-dock position, approximate planarization and alignment in the X, Y, and θZ directions has been achieved. Mating and planarization become more accurate as the mating members are actuated and the guide pins 912 are more fully inserted into their respective guide pin apertures 912a.

In general, the ready-to-drive position is one in which the alignment mechanisms in the two halves of the docking member have at least partially engaged and some, but not necessarily all, axis alignment has been achieved. For example, in docking members of the type described in the '815 patent and those of the type described in Figures 9(a) and (b) and 12(a) to (d), this is as stated above As described above, the tapered guide pin 912 has been further inserted into the guide pin aperture 912a such that the cam follower 1210 on the handler 1208 has been inserted into position in the docking cam 910. As such, test head 100 is typically aligned with the target device to within a few thousandths of an inch and within approximately 1 degree of coplanarity. As another example, in a docking member as described in US Pat. No. 5,982,182 (hereby incorporated by reference), this is the position where the moving contacts have engaged and the system is ready for final linear motion perpendicular to the target.

The solution to the above problem applies to the balancing system described herein in Figures 1(a) to 4(b) (described in U.S. Provisional Application 60/234,598 to Ny et al.), and also to the balance system described herein in Figure 6. BCAAs described in (a) to 9(b). These have several aspects in common; one is that the balance system is an integral part of the BCAA. Thus, BCAAs and/or balance systems can be added to the manipulator and test head 100 as required to achieve balance and compliance where needed or desired.

In this application, the test head 100 may be brought into and locked in a compliant neutral position prior to docking, as previously described. Just before the final docking in very close proximity to the docking system, the shaft or shafts can be balanced so that it or they are in a compliant state. Since this is accomplished by locking the lock 118, there is no risk of sudden, unpredictable, and potentially dangerous movement of the test head 100. Then, with the system in a compliant state, the test head 118 is undocked for a compliant motion to continue docking.

As noted earlier herein, pressure sensors ( 606 a , 606 b ) may be incorporated and arranged to measure the gas pressure on each inlet of the pneumatic cylinder 128 . A controller (not shown) receives a signal from the pressure sensor indicative of the measured pressure. This paper previously described the use of pressure sensors to restore a state of equilibrium prior to disengagement. Without this approach, normal operation would disengage the test head 100 with the balance system lock locked, which forgoes the opportunity for compliant movement during disengagement.

Pressure sensors (606a, 606b) may also be used in systems that maintain a balanced state throughout the time test head 100 is docked and as testing continues. To accomplish this task, the test head 100 is brought into a ready-to-dock position. All axes with a balance system or BCAA go into compliance readiness. That is, keep them balanced, and keep them in the pliable neutral zone. The pressure in the pneumatic cylinder 128 is then measured and recorded by the controller. Next, the latch 118 is released and the test head 100 is docked. Depending on the type and application of the docking system, the lock 118 may or may not be re-locked when the test head 100 is finally docked. While the test head 100 is docked, the controller constantly monitors the pressure in the pneumatic cylinder 128, compares them to recorded values, and manipulates the valve 604 so as to maintain the cylinder pressure substantially constant. This maintains a substantially constant force on the piston 130 and connecting rod 114 and keeps the system in the desired state of equilibrium. Clearly, the test head 100 can be disengaged for the compliant motion available in equilibrium.

More specifically, whether the docking system is latching or non-latching will determine the available modes of operation. The latch dock provides a choice between these two modes of operation as follows:

1. Maintain balance by disengaging lock 118 while docking test head 100 (as described above). By disengaging the lock 118 while disengaging, the opportunity for balance and compliant movement is maintained.

2. While docking the test head 100, by locking the locker 118, the balance is not maintained. For disengagement, there are two sub-choices:

a. Restoring balance prior to disengagement and preserving the opportunity for compliant movement by disengaging the lock 118 while disengaging.

b. While disengaging, no balance is restored and the lock 118 must remain locked. This method does not require a pressure sensor.

In non-latch docking systems, generally speaking, when the test head 100 is docked, the latch 118 needs to be locked. Balance may or may not be restored prior to disengagement, as required by the compliance movement.

Now consider operation in a system using a latch, actuator driven docking. This is the type of system that is most widely used today. The basic two-point docking device as described in U.S. Patent No. 4,589,815 can be generalized to include three One or four sets of guide pins 912, guide pin apertures 912a and cord drives. Such two-, three- and four-set mating members are widely used in the industry for this purpose. Although such devices may be manually actuated by the operator by applying force to the docking handle 914, other types of docking actuators including electric motors, electric or gas linear actuators, and/or vacuum operated devices are also known. . In general, the test head 100 will interface with a plug-in device handler, wafer inspector, or possibly other devices, which are collectively referred to by the term processor 1208 as "target devices." For docking, the following sequence can be followed:

As shown in Figure 15(m), the system is ready to move the test head or load 100 from the initial point of separation from the target device to the ready-to-dock position.

a. As shown in step 1620, the compliance actuators and/or BCAAs on the horizontal (in-out, side-to-side, and roll) axes enter a compliance-ready state and lock their locks 118 .

b. As shown in step 1621, put pitch and roll BCAA into compliance ready state and lock their locks 118 .

c. As shown in step 1622, the BCAA in the vertical drive mechanism enters the compliance ready state (using the compliance ready sequence) and unlocks its lock 118 .

d. As shown in step 1623, the test head or load 100 is now moved to a position ready for docking with the target device. To this end, the controller controls the actuators to move the test head 100 along the path of motion into a ready-to-dock position. Generally speaking, this is where the alignment mechanisms in the two halves of the mating piece have come into close proximity or initial contact, but not yet fully engaged. For example, in many types of docking pieces that have guide pins 912, it is that the docking pins 912 are already close to their mating pin apertures 912a in the target device or are just mating with their mating pin apertures in the target device 912a engaged position. The controller may use some algorithm to accomplish this, or the operator may direct the process by using buttons, joysticks, and/or other suitable input devices with the controller.

From ready to docked to fully docked

e. Preparation: All axes of the manipulator with BCAA are rebalanced. As shown in step 1624, each BCAA is brought to equilibrium in turn. Bringing a particular BCAA into balance may upset the balance of some other BCAAs. This process is then repeated until each of the BCAA's force sensors 120 indicates a state of equilibrium (typically, this is achieved by 3 or fewer iterations, usually 2 iterations).

f. As shown in step 1625, if a pressure sensor is included in any of the BCAAs, then the current pressure is read by the controller and stored in controller memory for later use.

g. Unlock all locks 118 as shown in step 1626, allowing balanced compliant motion along all axes (note that, by design, each When the motor is not activated and the roll axis is not being driven, the clutch of the roll axis motor is automatically disengaged). Now, the test head 100 or load is in a balanced compliant state and can be moved by an external device with typically less than 25 to 30 pounds of force.

h. Move the test head 100 further to the ready-to-drive position. This is an overall movement that is usually less than 1 inch, and this movement can and is usually done by hand. However, in more complex systems, a controller (not shown) may be used to force test head 100 into this ready-to-drive position. To do this, the controller forces the test head 100 to move along a straight path perpendicular to the plane of the processor (or other target device) electrical interface 1228 . select the appropriate axis or axes for motion (vertical axis for horizontal butt joints; in-out or left-right for vertical joints; or a combination of vertical and in-out or left-right for inclined butt joints), The lock 118 is locked and the actuator is used to force the test head 100 into position, thereby unlocking the lock 118 again (leaving the other shafts unlocked to allow the necessary compliant movement, as described above, this to enable alignment of the test head).

i. The docking actuator is now activated, drawing the test head 100 into full alignment with the target device 1201, as shown in step 1628. In a manually actuated docking piece such as the docking device shown in Figure 9(a), this is accomplished by applying a force to the docking handle 914 which causes the docking cam in Figure 9(a) to 910 or the rope drive (part of unit 914) in Figure 12(a) is rotated. In docks with powered actuators, this is achieved by appropriately energizing the actuators. As the test head 100 is drawn into the final docked position, it is free to compliantly move in all six spatial degrees of freedom. If desired, although this is not always preferred, selected axes may have their locks 118 engaged to restrict compliant motion in the respective degrees of freedom.

Typically, the docking mechanism is designed such that the test head 100 is effectively latched in place when the docking actuator has reached its fully docked limit. Thus, the test head 100 is now fully docked, latched in place, and mechanically ready to test the device.

Now consider operation in a manipulator-driven system with non-latch docking. In such systems, the manipulator must overcome the forces associated with docking. These forces typically result from the need to compress hundreds or thousands of spring-loaded pogo pins and/or to couple hundreds or thousands of mating contacts in an electrical connector. For each such pogo pin or contact requiring a few grams or ounces of force, a total mating force of several hundred pounds is often encountered. For manipulator driven docking, the manipulator drive mechanism used in the docking must be able to overcome the docking forces, as well as the forces required to move the test head or load 100 in free space. If the drive shaft is balanced, then the drive force can be reduced.

It is advantageous to use a BCAA mechanism in the shaft that is the drive shaft during manipulator driven docking. First, the drive shaft actuator can generally be easily sized to have sufficient drive capacity to overcome the docking force, and the force required to move the test head or load 100 . Furthermore, if the shaft under consideration is designed to move it in equilibrium with the lock 118 disengaged, then a relative position sensor can be used to detect whether an obstacle or obstacle is encountered (see, for example, the reference to Fig. 10(a ) to the discussion in 10(f)). In addition, the force sensor 120 may be used to detect obstacles and obstructions encountered while the test head 100 is being moved in the locked state of the lock 118 .

Recall that in actuator-driven docking, typically, the docking device is provided with guide pins 912 on one side and mating guide holes 912a on the other side to guide the test head 100 to the device handler (FIG. 12(a) 1208 in ) or the test point alignment of the detector. Two guide pins 912 fitted into two snug guide holes 912a will provide alignment in five spatial degrees of freedom, provided the guide pins 912 are of sufficient length and the guide holes 912a are of sufficient depth. If the mating surface is planar, these degrees of freedom include X, Y, and θ in-plane and pitch and roll relative to the plane. The remaining degrees of freedom are the vertical distances between the two planes controlled by the docking actuator mechanism. This technique can also be used in manipulator-driven docking; however, other techniques are known, such as that disclosed in [Graham et al.] discussed above.

The following outlines the procedure for manipulator-driven docking, which is shown in the flowchart shown in Fig. 15(n). This process follows the general form of the previous actuator-driven docking process.

1) The system is ready to move the test head or load 100 from the starting point of separation from the target device to the ready-to-dock position.

a) As shown in step 1630, make the compliance actuators and/or BCAAs on the horizontal (in-out, side-to-side, and roll) axes into a compliance-ready state and lock their lockers 118 .

b) As shown in step 1631, put pitch and roll BCAA into compliance ready state and lock their lockers 118.

c) As shown in step 1632, put the BCAA on the vertical drive mechanism into a compliant ready state (using the compliant ready sequence) and unlock its lock 118 .

2) As shown in step 1633, the test head or load 100 is now moved to a position ready for docking with the target device. To this end, the controller controls the actuators to move the test head 100 along the path of motion into a ready-to-dock position wherein the two halves of the docking device are already close to or just engaged with their counterparts. The controller may use some algorithm to do this, or the operator may direct the process by applying buttons, joysticks, and/or other suitable input devices to the controller.

3) From ready state to docked state to fully docked state

a) Preparation: Rebalance all axes of the manipulator with BCAA. As shown in step 1634, each BCAA is brought to equilibrium in turn. Bringing a particular BCAA into balance may upset the balance of some other BCAAs. This process is then repeated until each of the BCAA's force sensors 120 indicates a state of equilibrium (typically, this is achieved by 3 or fewer iterations, usually 2 iterations).

b) As shown in step 1635, if a pressure sensor is included in any of the BCAAs, then the current pressure is read by the controller and stored in controller memory for later use.

c) As shown in step 1636, the axes to be driven and controlled are determined. These may be predetermined by design in the specific equipment, but, in general, the equipment with which the test head 100 is being interfaced will determine them. Typically, the selected axis at least effects movement along a path perpendicular to the processor or detector interface plane, which may be horizontal, vertical, or inclined at an angle to the horizontal . In some applications, an axis that can be driven and controlled to enable planarization of two surfaces brought together is desirable.

d) As shown in step 1637, the lockers 118 in the non-drive and non-control axes are all released to allow balanced compliant movement along these axes while moving to the fully docked position. Holds the lock 118 of the axis being driven and controlled in the locked state (note that, as described in U.S. Patent Application 09/646,072 to Holt et al., by design, whenever the motor is not driven and the roll axis is not When driven, the clutch of the roll axis motor is automatically disengaged. Note also that if the roll axis is to be driven as part of a docking motion, it will need to be equipped with means for non-torque limited operation as described above). Now, the test head or load 100 is in a balanced compliance state.

e) As shown in step 1638, the test head 100 is further moved to the preliminary alignment position determined by the sensor and/or the tapered guide pin 912 is partially inserted into the guide pin receptacle 912a. In this way, test head 100 is aligned with the target device to within a few thousandths of an inch and within approximately 1 degree of coplanarity. This is an overall movement that is typically less than 1 inch, and, in a manipulator drive system, typically moves with a controller to force the test head 100 into this preliminary alignment position. To this end, the controller forces the test head 100 to move along a straight path perpendicular to the plane of the docking zone in the target device. If axes other than those selected above in step 2)c) are used in this step, then they must be locked before the move and unlocked again after the move (leaving the other axes unlocked state to allow the necessary compliance).

f) As shown in step 1639, the test head 100 is driven to a fully docked position with the target device. The axes selected above in step 1636 are driven and controlled by the system controller to move the test head 100 along the appropriate path to its fully docked position. As the test head 100 is driven to the final docking position, it is free to make compliant movements in all non-selected axes. However, some non-selectable axes may have their locks 118 engaged to restrict compliant motion within the respective degrees of freedom, if desired; however, in general, this is not the preferred mode of operation. When this movement occurs, monitor the appropriate position sensors. At least one or more position sensors are required to signal when the test head has reached its fully docked position. Provided the appropriate axes are driven and controlled during this process, other sensors may be used to maintain and improve the alignment and planarization of the test head 100 as necessary as the test head 100 is moved into the docked position.

4) As shown in step 1640, when the test head 100 is fully docked, the drive shaft (selected above in step 1636) is stopped and held in the fully docked state. Unless a latch mechanism is provided in the docking hardware to keep the test head 100 in its docked position, they must remain in this state during subsequent testing and use of the test head 100 . Such a latch may be controlled by a controller and activated upon a signal indicating that the docked position has been reached.

Alternative techniques for maintaining the test head in a docked position with the test head apparatus include locking or unlocking some or all of the manipulator motion axes. Another technique is to maintain the equilibrium state of the test head 100 by monitoring and maintaining the pressure in the cylinder 128 of the BCAA.

In the docking method, two intermediate positions of the test head 100 are identified in each case: the ready-to-dock position in both cases, and the ready-to-drive position in actuator-driven docking and the ready-to-drive position in manipulator-driven docking. The corresponding preliminary alignment positions in . In some cases, the two intermediate positions can be the same. Additionally, in certain dockings, a mechanism catch may be used to capture the test head 100 and hold the test head in either or both of these intermediate positions. When these catches are activated, they prevent the test head 100 from leaving the docked position obtained for the docked position, but instead cause it to move close to the docked position. Using these methods augments the methods above, rather than creating new ones.

While preferred embodiments of the present invention have been illustrated and described, it should be understood that such embodiments are provided by way of example only. Numerous variations, changes and substitutions may be made by those skilled in the art without departing from the spirit of the invention. Accordingly, we consider the appended claims to cover all such modifications as fall within the spirit and scope of the invention.

Claims (31)

1. An apparatus for supporting a load comprising: a force sensor for detecting a force received from said load originating from said load in an unbalanced state such that a torque is generated about a rotational axis of said load; A force source for providing a counter force relative to the load in a direction opposite to the force in response to the force detected by the force sensor. 2. The apparatus of claim 1, further comprising a lock that maintains the position of the load when the force sensor is detecting the force and when the lock is in a locked state. 3. The apparatus of claim 1 , wherein said force source provides compliance of motion to said load when said load is in balance with respect to said shaft and when said lock is in an disengaged state. scope. 4. An apparatus for balancing a load which is out of balance such that a torque is generated about an axis of rotation of the load, said apparatus comprising: a force sensor for measuring the amount of force exerted by the load on the force sensor due to the load imbalance; and A force generator for generating a reaction force relative to the load and in a direction opposite to the force to counteract the force. 5. The apparatus of claim 4, further comprising a lock that holds the load in place when the force sensor is measuring the magnitude of the force and when the lock is in a locked state. Location. 6. The apparatus of claim 5, wherein the force generator provides a compliant range of motion to the load when the load is balanced about the shaft and when the lock is in the disengaged state . 7. A method of balancing a load about an axis, said method comprising the steps of: measuring at least a portion of the force exerted by the load due to the load imbalance; and A counter force is applied relative to the load and in an opposite direction to the force to counteract the force so that the load is closer to equilibrium than before the counter force was applied. 8. The method of claim 7, further comprising the steps of: During the measuring step, the load is locked in position. 9. An apparatus for supporting a load, said apparatus comprising: a force sensor for detecting a force received from the load resulting from an imbalance of the load relative to the load's axis; a force source for providing a reaction force relative to the load and in a direction opposite to the force in response to the force detected by the force sensor; a compliant unit for providing the load with a range of motion relative to the shaft, the compliant unit including a lock that secures the lock in the locked state when the lock is in a locked state position within range of motion; and A drive unit coupled with the compliance unit is used to drive the load along the direction of the shaft when the locker is in the locked state. 10. The apparatus according to claim 9, wherein when the locker is in an unlocked state, the position of the locker within the range of motion is changeable as the drive unit drives the locker . 11. The apparatus of claim 9, wherein said axis of said load is a rotational axis. 12. The apparatus of claim 9, wherein the compliance unit includes stops at each end of the range of motion, the lock being between the stops. 13. The apparatus as claimed in claim 12, wherein, when the locker is in contact with one of the stoppers, the drive unit moves along the shaft with the locker in the disengaged state. direction to drive the load. 14. The apparatus of claim 9, wherein the drive unit is powered. 15. An apparatus for supporting a load, said apparatus comprising: a force sensor for detecting a force received from the load resulting from an imbalance of the load relative to the load's axis; a force source for providing a reaction force relative to the load and in a direction opposite to the force in response to the force detected by the force sensor; a compliant unit for providing the load with a range of motion relative to the shaft; and A drive unit controlled independently of the force source, coupled to the compliance unit, for driving the load in the direction of the axis. 16. The apparatus of claim 15, wherein the shaft is a rotating shaft. 17. The device of claim 15, further comprising: The locking device is used for fixing the position of the range of motion shown when the locking device is in a locked state. 18. The apparatus of claim 15, wherein the compliance unit includes a stop at each end of the range of motion. 19. The apparatus of claim 15, wherein the drive unit is powered. 20. An apparatus for supporting a load in a direction, said apparatus comprising: a force sensor for detecting a force in said direction resulting from said load being driven in said direction; a force source for providing a reaction force relative to the load and in a direction opposite to the force in response to the force detected by the force sensor; a compliant unit for providing a range of motion to the load in the direction, the compliant unit including a lock that fixes the lock in the range of motion when the lock is in a locked state within the location; and A drive unit coupled to the compliance unit and controlled independently of the force source for driving the load in the direction when the locker is in the locked state. 21. The apparatus of claim 20, wherein said position of said lock within said range of motion is actuatable by said drive unit when said lock is in an unlocked state. And change. 22. The apparatus of claim 20, wherein the lock is within the range of motion when the lock is in the disengaged state and when the load is substantially weightless in the direction The position of can be changed by driving the locker by the drive unit. 23. The apparatus of claim 20, wherein the compliance unit includes a stop at each end of the range of motion. 24. The apparatus of claim 20, wherein the drive unit is powered. 25. An apparatus for manipulating a load in a vertical direction, the apparatus comprising: a force sensor for detecting a force along said vertical direction resulting from said load being driven along said direction; a force source for providing a reaction force relative to the load and in a direction opposite to the force in response to the force detected by the force sensor; a compliant unit for providing a range of motion to the load in the vertical direction, the compliant unit comprising a locker, when the locker is in a locked state, the locker fixes the locker during the movement location within range; and A drive unit controlled independently of the force source for driving the load in the vertical direction when the locker is in the locked state. 26. The apparatus of claim 25, wherein the compliance unit includes a stop at each end of the range of motion. The locker is between the detents. 27. The apparatus according to claim 26, wherein, when the locker is in contact with one of the stoppers, the drive unit moves in the vertical direction with the locker in the disengaged state. drive the load. 28. The apparatus of claim 25, wherein the drive unit is powered. 29. A method of interfacing an electronic test head secured in a test head manipulator with an electronic device processor, the method comprising the steps of: measuring the magnitude of an unbalanced force along or about at least one of a plurality of axes of motion of the test head manipulator; and A counter force in a direction opposite to the unbalanced force is provided for the unbalanced force until the unbalanced force is substantially zero. 30. A method of interfacing an electronic test head with an electronic device processor, the method comprising the steps of: measuring a magnitude of an unbalanced force along or about at least one of a plurality of axes of motion of the test head manipulator; providing a reaction force to the unbalanced force in a direction opposite to the unbalanced force until the unbalanced force is substantially zero; and While applying the counter force, the test head is moved toward a docked position relative to the device handler. 31. A method of interfacing an electronic test head secured in a test head manipulator with an electronic device processor, the method comprising the steps of: measuring a magnitude of an unbalanced force along or about at least one of a plurality of axes of motion of the test head manipulator; providing said unbalanced force with a reaction force in a direction opposite to said unbalanced force until said unbalanced force is substantially zero; and While applying the counter force, the test head is moved toward a docked position relative to the device handler.

Images